Upgrade Progress
6/13/2024
working on Chuck's machine
dropped all jwars tables OK
ran createJwarsTables.bat OK
ran xyzzy.bat OK
ran disableContraints.bat OK
ran impDumpFile.bat OK
ran enableContraints.bat OK
VP runs
Several sims run.
6/12/2024
create hpac iso - can't read disk
install hpac
test dlls that I think are missing
walkbacks are in c:\ProgramData\Jas\walkbacks. This is an ini file setting.
6/10/2024
fixed date and time batch code to account for month, day, hour, and minute being less than 10.
Apparently successful Oracle load
run createJwarsTables.bat
run disableConstraints.bat
run impdp
run enableConstraints.bat
run showBadIndexes.bat
all good
To start over, repeatedly run dropJwarsTables,bat and selectJwarsTables until the log for the latter shows no results.
Next is NSIS
6/9/2024
used system restore to get a clean copy of win 10.
installed win 10 and took restore point
install oracle and use snipping tool to save screen shots for later use in document.
installed oracle and took restore point.
backed and replaced tnsnames.ora, sqlnet.ora, and listener.ora.
created table spaces by running createTb11g.bat
run the 4 batch files in createUserAndRoles in 1-4 order.
changed JWARS password to password (lower case).
Oracle converts jwars to JWARS when using sqlplus.
Many many problems with createJwarsTables.sql
TABLESPACE “JWARS_PRIMARY”;
DEFAULT));
TABLESPACE “JWARS_ENVIRONMENT”;
TABLESPACE “JWARS_SCENARIOCOMPONENT”;
Finally got it to work.
did an impdp of latest .dmp file
had parent key not found problems with ENV_STATIC_GRID, ENV_SEASONAL_GRID, JAR_TABLE_COLUMN, and JAR_TABLE_STRUCTURE because of loading order.
impdp also loaded all J99999 run outputs which are hard to delete.
deleteAllJwarsTables.sql is a list that does not l=know about them.
solution to both of the above is to have a parfile that specifies tables and order.
need either a 'delete all where owner = jwars or jut drop the table spaces and re-run createJwarsTables.bat.
The load of env tables is envArea, envScenario, envScenarioParam, envScenarioParamLayer, envScenarioGrid, envStaticGrid so delete in reverse order?
The load of jar* data is jar_table. jar_table_column, jar_table_structure, so delete in reverse order.
The above two could be a second and third import using explicit table names
6/8/2024
kill failed instrument output DONE
created HowToBackupOracle.txt in Jas/Oracle/backup DONE
exp oracle using these directions DONE
install oracle using oracle.doc
6/7/2024
createJwarsTableCreationSql was created by the Smalltalk script z:\avast\workspaces\tableCreationSql..
It produces a .log and a .lst (Oracle) file with CREATE TABLE defs for the Jas tables.
There are lines where one ends in DE and the next line is FAULT. These must be combined.
To combine these lines, use Notepad++ to search DE\r\n and replace it with DE in either the log or the lst file.
-
6/3/2024
JWARS_PREFERENCES - Chuck 200 Dev 198
SCENARIO Chuck 567 Dev 568
SCENARIO_COMPONENT - Chuck 11375 Dev 11372
Get everything working on Dev then clone all to Paul and Oracle to Chuck
where are walkback logs?
install HPAC
run defs all use /users/chuck as instrument dir
what's the difference between a regular and an open run def?
how does export create a directory name? should be date time and run def?
5/24/2024
Jas work was completed in spring 2022.
Created Oracle files on Chuck's machine 5/1/2022.
Jas HCI blows up reading JWARS_Preferences where owner = 'JAS' and window = 'Run Output Browser' and category = 'VP Launch Export Default' because of a mix of old and new ObjectDumper formats. How did this happen.
This happened on Chuck's machine and my Jas Dev machine.
Wrote two sorta Sunit tests - one to count all 7 tables that use JwModel and the other to read them.
Count showed that there were 203 records in JWARS_Preferences.
The read blew up on this bad record.
I deleted it. Sunit read ran fine after that.
Then dumped the JWARS_Preferences table and got 202 rows.
Dropped the table.
Then imported JWARS_Preferences for the exp2022_04.dmp file and got 198 rows all of which read ok.
VP Launch Export Default is not among them.
Trying to read all 11k scenario_components blows up on #4932 because it can't find the JwChemBioMaterialLiquid class.
This ^^ is a completely different problem than JWARS_Preferences, which is probably not worth trying to fix because there are no runnable scenarios that use this component..
To do: count all tables on both Chuck's and Jas Dev. DONE
Chuck's machine and Jas Dev have the same hci, sim env and jacs icx files..
1/24/2024
most vp code is in app ExtVideoPlaybackApp
vp browser is VideoPlaybackBrowser
do VideoPlaybackBrowser open
look for method containing vpDir to see how directories are selected.
VpGrouping objects show scenario names and hold runs in VPRecordLocatorFile objects
listDisplayLabel is the printString for a listBox
modified VPRecordLocator»listDisplayLabel to make VP show instrument directory name instead of the runId
5/1/2022
4/30/2022
fix env load, reduce, and packaging rule errors
got jwars cannot create seeeion in new jas machine
copied Z:\jas\JSO\Releases\R2.0 SR3.20\R20 SR320 JAS Database & Source\Database\new.
change long raw to blob and number(38) to number.
ran it and all appears wll.
but got many *
ORA-31693: Table data object "JWARS"."ENV_SCENARIO_GRID" failed to load/unload and is being skipped due to error:
ORA-29913: error in executing ODCIEXTTABLEFETCH callout
ORA-02291: integrity constraint (JWARS.SYS_C0018098) violated - parent key not found
4/29/2022
chuck's new machine
could not find xml because the path in vmcfg.ini pointed to an Instantiations install Fixed in new vmcfg.ini
blew up looking for ws2_32 because I addded progra~2\jas to the path instead of Program Files(x86)\jas Fixed
todo
fix vmcfg.ini to remove dev stanzas, use proper path for xml, and comment Fixed
remove old ws2_32, dated 2010, and test on reference machine Fixed
do we need gdk.cat? In new nls directory
whJacs and whEnv are older versions of abtntsrv.exe (2018) Fixed
need nls in programData and add to ini Fixed in new vmcfg.ini
4/27/2022
used the dropTn script to drop the table spaces
then stopped Oracle and deleted the files
ran createTb
wrote new create1PasswordExpire to make the password on the DEFAULT profile never expire
ran it and the 2, 3, 4 scripts
imported the 422 dump file and still had index errors
wrote showBadindexes and found 14
found an old index fix script
ran it and showBadIndexes found nothing wrong.
4/26/2022
To do
look at errors in the import log. DONE
need to drop all indices and reindex everything. DONE
find all profiles and look at password expire. DONE
4/24/2022
added errorLogFile setting in jasS.ini
ran alter table env_scenario_grid modify(scenario_grid blob)
ran for about 25 minutes and packaged sim from hci is still good
4/17/2022
4/14/2022
Got
AbtError: rc=-2 for 604 in an AbtOracle10DatabaseConnection at (4/12/2022 6:57:59 PM)
* 'ORA-00604: error occurred at recursive SQL level 1
* ORA-01000: maximum open cursors exceeded
* [JwInstrumentBuffer>>#privilegedStoreToDatabase]
4/12/2022
big problem in sim xd was * dnu loadShadows
code ran fine in jso.
finally looked in problems.es and found that loadShadows was not in the sim XD code
it was in JwgAbstractStorage which was in JSO but not in sim XD
copied JwgAbstractStorage»loadShadows to JwAbstractStorage»loadShadowsForActualObject
10 days of AJH TA Sea Viking runs 6000:1 and 7000+:1 running silently
4/11/2022
try to go back to last good (?) non XD
use non XD sim rules R3.0 SR2.0 1.0
use Sim Base R3.0 SR1.02 1.2 - dated 1/14/2022
use Pkg Instrcutions R3.0 SR1.02 1.2 - no date
get the Windows CommonControls failed to init message which is just a warning but not until I kill the batch file. get nothing from the jacs.
wrote workspace script to write method applicationNamesTOpackage for all app loaded by the XD map
actualIncludedSubApplicationNames went from 412 on 1/15/22 to 526 on 1/17/22 when it first reported to 537 on 4/11/22
current results of -report were from the batch script runWhSim.bat which will run as non Admin.
Jacs never show a response from the same command
task manager does not show them hanging,; they probably never start.
4/10/2022
4/9/2022
used a workspace to sort all the names in actulLoadedSubApplications
then created doNotReduce rules for them all.
still getting LookupTable does not understand shadows.
last JSO JwiNonXdSimNTHeadless was R2.0 Sr3.16
301 apps to package. 555 actual included
latest JSO JwiNonXdSimNTHeadless was R3.0 SR2.0 1.1
111 apps to package. 580 actual included
there is no real difference between jso from Sac and my latest except for stuff like time zone
reduced runtime image (non XD) alwyays includes Windows maps, e.g., CommonGraphics
there is no easy way around this short of editing packaging instruction which may get overwritten by a Save
4/8/2022
xdSim dies on LookupTable DNU loadShadows even after adding do not reduce CLDT to packagins instructions
not clear where I got/copied the original xdSim packagina instructions
nonXD has 111 applicationNamesToPackage. xdSim has 120
nonXD has 516 actualIncludedSubApplicationNames. xdSim has 580
xdSim instruction say doNotReduce RcMath, but it was excluded by reduction.
the difference between Xd and notXd instructions is that the xd instructions say #RcMath not #'RcMath'
this ^^ is wrong. Both #abc and #'Def' work
4/7/2022
JwvSinRunInfo is in a sub app of JwvDatabaseModelsApp
JwvDatabaseModelsApp is in config maps Jas Dev JSO, VP, HCI, Sim XD, and Jads but not Jacs
latest version is R3.0 SR1.02 5.19 which is in HCI, JSO, and Sim XD.
the dataObject is a BLOB, not a LONG RAW, so should be ok.
latest version is R3.0 SR1.02 5.21 which is noe in HCI, JSO, and Sim XD.
new launch: method in Jas Dev JSO 1.8 works with wifi on!
4/6/2022
jacs is now starting the sim but it dies trying to read JWARS_SimRunInfo in retrieveSimRunInfo:
the gdkTimeStamp passed in looks ok and there is sometimes a record in ORacle and sometimes not
the error message is JAS Simulation System:ERROR:0:'?' and the result is an empty OrderedCollection
I do not think jacs is at fault
To do:
Fixed
HCI was locking the record when it wrote the BLOB
Sim could not read/find it
Add commitUnitOfWork after HCI read/lock
3/8/2022
I think the previous good jacs was 3/2
None of the maps from 1.36 on even start the sim and then have it fail the way it used to.
They all put it in scheduled with no warning message.
The Run Def gui show hostname DellWin* as the only available host
This ^^ is done in Jacs and not using messages.
images by hosts shows 127* with no tcp.
3/7/2022
sim reportConfig is getting sst errors because it has no ip address
last good sim was 1.3
1.4 DNU 'JwSimulationSystem does not understand configMapVersion'
1.5 GdkNetworking class»#packagingRulesFor: - Execution Error
1.6 gets the sst error
repackaged 1.3 and it reports ok
reporting problem was new version of JwrSimConfig which I created using Refactoring Browser.
Added just the new launch: method which still failed with all networking off.
Jacs found Vmware adapters until I uninstalled Vmware.
GdkSystem»localIpAddress fails because SciSocketAddress determineInternetHost: self localHostName gets an SciError in VAST 9.2 when not networking on. Fixed
3/6/2022
tried wireshark and it did not capture any 127 traffic. not sure why.
jacs is OK but has crappy erroe messages
went back to an xd sim dated 1/15/2022 and it does report to jacs
launch: method was still bad and sending Jacs addresses like DellWin10
need to fix method with text 'host missing' to say that sim will not run.
need jacs test and error message if there are no sims availab
receiveSimulationAddress: aString forSimulation: aNumber ←———-
3/3/2022
3/2/2022
repackage HCI and try to run sim
jasDevJso has a later version JwrJacsBaseApp then does pkgSim Fixed
from simvm.log: JwrSimImageConfig does not understand configMapVersionName Fixed
restored logDebug messages in JwvJACSModulesApp
JwvJacsServer»receiveSimulationConfigReport: never gets the reportConfiguration sent by whSim
3/1/2022
fixed jasDevJso jas.ini to look for libmi in progData\jas\dlls
IwedaPath=C:\ProgramData\Jas\iweda in jas for jasDevJso
same as above (2) for jasS.ini
fixed EffectsServerManager»loadParameters to exit if iweda not found
air and vp run defs now run for one day and auto select sim version.
JwgRunControlWindow show Connecting but nothing else with all messages on in VP. WORKS NOW
update GdkKernel in all maps DONE
need showBusyCursor in Run output browser NO
DK allOracle and alloracleXD are ok
jasDevJso loaded in an open image runs and outputs instruments
jasDevJso and jasPkgHci differ on JwgJwarsCore 6.09 vs 6.11 respectively
jasDevJso and jasPkgHci differ on JwgHciSystem 3.78d vs 3.78f respectively
update jso with these two newer maps and run
jasDevJso runs vp and air ok DONE
1/21/2022
fix missing slac and iweda message in hci.
test missing slac and missing iweda
test no hosts available DONE
fix missing environment message
1/20/2022
1/19/2022
1/18/2022
After successfully creating a sim, JwvJACSManager»scheduleSimulationsForDeveloperOn:hciSystemAddress: aSstEndpoint adds it to scheduledSimulations
Note: running JwvJacsServer»answerSimulationOn: asDeveloper: with developerFlag set true does not work
added many logDebug: to an oepn image jacs.
ran it with JwvJacs new launch
could not set breakpoints - got gccs errors
jacs says that a sim launch has failed if the sim never sends an update of some time value back to the jacs
this value will obviously never get update if the sim never starts.
task manager shows:
vast runtime 5.9M - environments
vast dev 44.9M - jasDevJacs open image
vast dev 334.9M - packaged HCI
It never shows another Vast exe running
A Host is a machine such as 'Dell6420Win10'
18 Jan 2022 21 35 56 | Debug : imagesByHost
18 Jan 2022 21 35 56 | Debug : Host '10.0.0.76'
18 Jan 2022 21 35 56 | Debug : imageName .\whSIM.icx
cfgVersion R2.0 SR3.20
cfgTimeStamp
packageTimeStamp
queryTimeStamp
logDebug this^ wgeb jacs gets report
fix other parameters and make names meaningful
fix “not simmm for xxx' message
1/17/2022
esvio.exe -report -j10.0.0.76:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIMNonXF.icx -nosplash does not report, and does not finish on a non-XD image in either Jacs or a cmd window.
both non/XD sim packaging were missing ome or more time zone apps and both rules.
repackaged xd.
repackages jacs to use esvio to run sims. -report works ok. sims fail.
no joy on comctl32.dll
1/16/2022
fix No DATUM function provided to libGCS - WGS84 DATUM assumed
blue window is JwgSimulationMonitor
Run Control window is JwgRunControlWindow
cmSinExec* kicks off sim
comctl32 check happens in WindowsPlatformWidgets class»initializeCommonControls.
Install the missing comctl32.dll file in the proper folder on the computer. If the operating system is 64-bit Windows, copy comctl32.dll to the “C:\Windows\SysWOW64” folder. If the operating system is 32-bit Windows, copy comctl32.dll to the “C:\Windows\System32” folder.
whSIM.exe -report -j10.0.0.76:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIMNonXF.icx -nosplash gets common controls error, does not report, and does not end when run from jacs.
It does finish but does not report when run from admin cmd window.
Adding comctl32.dll to jacs dir and progFilesx86\jas makes no difference.
nodialog does the same thing.
whSIM.exe -report -j10.0.0.76:8000 -lsimvm.log -ini:vmcfg.ini -i.\whsimxd.icx -nosplash does report and does finish when run from jacs.
run from admin cmd window is different
What is vast runtime path?
1/15/2022
created map Distributed Kernel - All Oracle XD
created map JAS Packaging: Sim XD
merged JAS Packaging: Sim Base and JAS Packaging: Sim C4I Extensions into it.
removed map JAS Packaging: Sim JADS
map loads
much fooling around to get image to package
was getting packagingRule does not compile because SSTInvocationHandler was missing
solution is to always load map in Dev then in XD
sim xd packaging has many jas warnings.
compare them to old sim packaging.
output image takes 4 minutes
esvio -report WORKS!
[Resources]
MILibrary=C:\temp\libmi.dll
SlacLibrary=C:\temp\libslac.dll
1/14/2022
the entry point for the packaged sim is JwSimulationSystem runHeadless.
add GdkSystem»logTTY: method to output runtime debug information
add debug messages to GdkSystem»launch
latest jas dev jso loads and runs
packaged sim ok, but running it with -report complains about Windows CommonControls failed to initialize properly because it cannot find comctl32.dll, but it continues to run and must be killed by the task manager.
tried removing all cg apps with no luck.
find what dll is missing and 1) copy it locally or 2) put it on the path
crtstdll.dll is an obsolete SSL dll that is no longer distributed by Instantiations. It is not the problem.
There is a comctl32.dll in C:\Windows\System32.
C:\Windows\System32 is on the Path
control is loaded WindowsPlatformWidgets»initializeCommonControls
whsim.exe is copy of nodialog.exe.
copied nodialog.exe.manifest to whSim.exe.manifest.
then from an admin cmd windows ran copy whSim.exe.manifest+
this “touches” the file's time date.
windows picks it up and Windows CommonControls failed is gone
BUT whSim.icx does not report and continues to run.
1/13/2022
revert to 2009 version of JwvJACSOSInterface»#startSimulation:launchDelay: because of many sdfs.
try running jasExp and new packaged hci against 7.5 jacs.
The packaged HCI starts then stops a sim on both the old and new jacs. no blue screen.
jasEXp gets requests sim execution but never gets a start then this sst warning and error old jacs
13 Jan 2022 11 52 10 | Trace : * [Request] SUBSCRIBE - **SimulationCreated** [USER - JAS@DELL6420WIN10] @ 127.0.0.1:51336
13 Jan 2022 11 52 10 | Trace : * UNIQUE RUN ID requested - JAS
13 Jan 2022 11 52 11 | Trace : * [Request] CONFIGURATION RUN HISTORY ENABLED [USER - JAS@DELL6420WIN10] @ 127.0.0.1:51336
13 Jan 2022 11 52 11 | Trace : * SIMULATION EXECUTION requested from HCI - JAS
13 Jan 2022 11 52 14 | Trace : Client [JAS Simulation System] connected from 127.0.0.1:51345
13 Jan 2022 11 52 14 | Trace : * [Update] CLIENT HEARTBEAT RATE from @ 127.0.0.1:51345
13 Jan 2022 11 52 15 | Trace : * [Update] SCHEDULED SIMULATION ADDRESS for 3819527531562003 [Sim Server - 3132:372e:302e:302e:3100:::51353] from @ 127.0.0.1:51345
13 Jan 2022 11 52 22 | Debug : [Server Heartbeat] Performing maintenance activities.
13 Jan 2022 11 52 22 | Debug : [Server Heartbeat] Sleep for 60 seconds.
13 Jan 2022 11 52 45 | Trace : WRN (SST): Handling <SstEndpointUnreachableError('notification timed out')> when calling sendSimulationCreatedAt:notifying: @ 127.0.0.1:51336
13 Jan 2022 11 52 45 | Trace : [STALE] User Client -> 127.0.0.1:51336. Last Outgoing Traffic (13 Jan 2022 11 52 15.375003), Heartbeat Rate (60)
13 Jan 2022 11 52 45 | Error : ERR (SST): SstRemoteMarshalingError('(''Primitive failed in: OSPtr>>#uint8At: due to Invalid offset in receiver'')') @ 127.0.0.1:51345 [ Server sending method --> sendInitializationSimInfo:to:]
jasExp runs angie air ok with new jacs. still get sst warning and no run control,
The packaged hci is attempting to run 321.icx which we do not know if it is any good,
verified that the packaged 3.20 I am running is from 9/30/2010. Verified
commented out most logDebug: but tagged them with jacsDebug
whSIM.exe -q3819541008531003 -uJAS -i.\whSIM.R20.SR320.icx -j10.0.0.76:8000 -e10.0.0.76:8002 -mn50000000 -mo560000000 -mi20000000 -ml20000000 -lsimvm.log -ini:vmcfg.ini -nosplash
running this ^^^^from an admin cmd windows shows Client [JAS Simulation System] connected from 10.0.0.76:52245 in jacs.log but does not run.
run from hci starts then stops right away
I have been doing apples to oranges comparisons all week.
the jasExp open image is starting a sim via jacs proxy not the jacs and is not running the packaged 7.5 320 icx but rather the 9.2 sim code in the open image. this is not the same code as the packaged 7.5 320 icx.
the new packaged hci uses either jacs to try to run the packaged 7.5 320 icx using an os call. the call from the jacs appears to work (jacs log says sims start) but then the sim stops right away.
the 9.2 jacs can run the packaged 7.5 320 image using os calls if it runs it with the -report option
I think the jacs is ok and probably the packaged hci too.
try running the 320 packaged image with a 7.5 whSim and both jacs. NO. this works
then create a 9.2 packaged sim and repeat with TranscriptTTY calls if necessary
1/12/2022
try packaging the hci as 320
^^^^ did not work
try to run open image EXP on different Use Sim For HCI options
^^^ can't select version in open image
need to fix packagingClassName first. DONE
jacs startup is JwvJacsNtServiceApp starting
code is JwvJACS new launch
jasc starts to launch in open image but missing jas.ini
added mucho logDebug: to jacs in latest config
1/11/2022
start 7.5 jacs in programdata\jas\jacs
start an admin cmd window
cd to C:\Program Files (x86)\Jas75\bin
run nodialog.exe -report -j127.0.0.1:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIM.R20.SR320.icx -nosplash -srvWorkingDir C:\ProgramData\Jas\jacs
cannot find the sim ics probably because srvWorkingDir is for services only
copy whSIM.R20.SR320.icx and jasS.ini to C:\Program Files (x86)\Jas75\bin
run nodialog.exe -report -j127.0.0.1:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIM.R20.SR320.icx -nosplash
simvm.log is created in C:\Program Files (x86)\Jas75\bin.
the first part of any simvm.log entry is info about the nodialog version, in this case 7.5
This proves(?) that a 7.5 nodialog can query a 2010 sim and report to a 7.5 jacs
cd to C:\ProgramData\Jas\jacs
ran it again in the cmd window and got results in programdata\jas\jacs\simvm.log
The result showd nodialog version to be 9.2 in this case
this proves that a 9.2 nodialog can query a 2010 image and report to a 7.5 jacs.
Problem was that the 9.3 whSim.exe was a copy of abtntsrv.exe
Should have been a copy of nodialog.exe.
Everything works using latest Jacs Oracle after the correct copy was done.
no op JwvJACSServer»serverHeartbeat
hci does start sims even when 3.20 is selected in the run def.
1/10/2022
JwvJACSOSInterface»reportSimConfigsInDirectory:toServer: looks in the current working directory (C:\ProgramData\Jas\jacs) for sim images using pattern '*hSIM*.icx' and writeS to the jacs.log:
Found .\whSIM.R20.SR320.icx in directory C:\ProgramData\Jas\jacs
Then it asks each image found to report its configuration by executing something like this:
whSIM.exe -report -j10.0.0.76:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIM.R20.SR320.icx -nosplash
The results should be written to simvm.log
VA Smalltalk, Version 9.1
(g)VM Timestamp: 4.0, 11/21/19 (101)
(C) Copyright Instantiations 1994, 2018. All rights reserved.
(C) Copyright International Business Machines Corp. 1994, 2006. All rights reserved.
Virtual Machine PID: 6224
VM Options {
newSpaceSize set to 2097152
}
Commandline args {
-report
-j10.0.0.76:8000
-lsimvm.log
-ini:vmcfg.ini
-i.\whSIM.R20.SR320.icx
-nosplash
}
** Reported configuration to JACS
Image Name - .\whSIM.R20.SR320.icx
Version Name - R2.0 SR3.20
Package Date - 9/30/2010 10:26:54 AM.000000
The latest versions of Jacs do not write anything to simvm.log and imagesByHost in JwvSimImageConfigManager is empty so there are no hosts to run sims.
BUT, the above slice from simvm.log dated 12/27/2021 suggests that this did work recently.
It does not explain why a 7.5 Jacs did not write to simvm.log.
Just packaged Loaded JAS Packaging: Jacs R3.0 SR1.00c/d/e/f/g/h. Runs but no log.
C:\ProgramData\Jas\jacs>whSIM.exe -report -j10.0.0.76:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIM.R20.SR320.icx -nosplash
StartServiceCtrlDispatcher error: (0x427) The service process could not connect to the service controller.
The above is Win10 with Admin Cmd window.
Runs fine on 7.5 and writes simvm.log
The problem may be with different versions of SST files.
A 7.5 Jacs talking to a 3.20 (2010) image probably is using the same SST maps.
But a Win10 Jacs may have a mismatch in SST talking to 3.20 (2010) images that does not work but shows no error.
Make HCI sim version be R3.0 SR 1.02 with no trailing version such as R3.0 SR 1.02 1.0
Config maps can have the same name.
Package the sim using R3.0 SR 1.02 as map name and openImageVersionName
1/9/2022
created Distributed Kernel All Oracle.
replaced existing DK maps with it in VP and HCI, both of which package
Jacs uses Distributed Kernel Jacs Oracle
HCI is complaining about transforms
1/8/2022
created Distributed Kernel Jacs Oracle and Jas Packaging: Jacs Oracle.
Both load and work.
Spent hours trying to make JwvJACSOSInterface»reportSimConfigsInDirectory:toServer: print more info.
Was getting sdf files (xd walkbacks) so reverted to 2009 version
Jacs working dir is ProgData\jas\jacs
Jacs uses the path to find whSim.exe
purged vast 10 config maps
1/7/2022
1/6/2022
1/5/2022
1/3/2022
12/28/2021
update dist kernel oracle to get lastes oracle
AbtDbmOracle10BaseApp»ociLobRead:amount:piece has two errors:
fix this (from jacs.log) in jas.ini(?) The setting {DatabaseType} is missing under the section {Database}.
Jas Dev JSO could not have worked because it only had 1 of the 2 ociLobRead: fixes in it.
Sim never seemed to start until I loaded latest AbtOracleLobBuffer and set defaultBufferSize to 1000000
JasDevEXP runs!
look at editions of AbtDbmOracle10BaseApp and get one that does not have Vast 10 code in it
fix wtf is going on with VisualWorksCLDTCompatibility
update GdkKernel (help→About) after everything works
jso3 runs. use it as a base going forward.
12/27/2021
packaged hci ok using open map from 11/5/21
packaged image blows up looking for Oracle 7
the two hci.exe from april 2021 also blows up looking for oracle 7
I am using older versions of renamed abt.exe and abtntserv.exe progfiles/jas
abtntserve should be whAudit, whEnv, whJads, and whJacs.exe Fixed
abt.exe should be wHci and wVp.exe Fixed
* hci reduction takes 3 minutes
hci output takes 5 minutes for a total of 7 minutes
fixing the exes (see above) make the oracle 7 errors go away.
now get
12/24/2021
12/23/2021
12/22/2021
used env devVP image and VP runs in an open image.
put mapInfoDir in the videoplayback.ini file under Genera;. use cwd\mapInfo if no override
warn users of missing files using FQDN name
jas dev JSO maps do not load in devJso
lasted map in 11/5, but I did runs on 11/6 and 11/7 but do not know which image and map.
found 9.2×64 under \code which might have been deleted. copied all builds to synshare
11/7/2021
jas dev vp map loads
VideoPlaybackBrowseer open starts VP
need to find videoplyback.ini file
the mapinfo directory is hard-coded as'.\MapInfo\frames.dat'.
videoplayback.ini is in C:\ProgramData\Jas\client.
it contains no override location for mapinfo
packaged vp and got wVp.icx of 15,963.
previous versions were 17,413
11/6/2021
Honor rolll is Ctrl J from launcher
JwarsAboutDialog gets the package time stamp from hciSystem packageTimeStamp in method normalText
GdkSystem class»packageTimeStamp is replaced by runtimePackageTimeStamp in a packaged image.
11/5/2021
we are using an Abt10Cursor
run with all instruments still blows up even when bufferSize = 1,000,000.
11/4/2021
changed gdkOracleDataType for Object(blob) and Integer (number19).
fixed JwvSimAuditInfo insert:ifError and not get instrument output for a VP run.
Getting Ora-01002 to many open cursors with all instruments selected
changed number(38) in descriptor_definition and descriptor_associations tables;
ran AJH with 48 air instruments ok, but get ora-01000 running with all
11/3/2021
commented out the writing of JwvSimAuditInfo.
scenario with all and just vp instruments runs to completion but no instrument output.
cannot 'describe' the J202111… table to see what is in it.
sql to create the JwvSimAuditInfo table which is the instrument table.
Audit_object is a JwvRawString. Making it a JwDataObject still gets a LONG RAW data type.
CREATE TABLE JWARS.J20211103161947734002_1 (Audit_Object long raw,
Begin_Record_Time NUMBER, End_Record_Time NUMBER, Instrument_Category varchar2(255),
Instrument_Count number(38), Instrument_Name varchar2(255), Long_Raw_Size number(38),
Replication_Id number(38), Run_Id char(20), Sequencer_Id number(38), Simulation_Time varchar2(20),
etc.
data types for columns are defined by gdkOracleDataType methods.
Object has long raw. should be blob.
Number gets called with a size.
Integer call it with 38. Changed to 19.
sql to create the JwvSimAuditInfo tableis now this:
CREATE TABLE JWARS.J20211103173620343001_1 (Audit_Object blob,
Begin_Record_Time NUMBER, End_Record_Time NUMBER, Instrument_Category varchar2(255),
Instrument_Count number(19), Instrument_Name varchar2(255), Long_Raw_Size number(19),
Replication_Id number(19), Run_Id char(20), Sequencer_Id number(19), Simulation_Time varchar2(20),
etc.
reverted to JwvRunOutputModelsApp - R2.0 SR2.53Q 3.33.0.0 - (6/22/2010 4:26:22 PM)
all later versions have odbc. ← WRONG
went back to JwvRunOutputModelsApp - R2.0 SR3.15 3.43 - (8/31/2010 11:11:25 AM) which has many more instance variables that are being passed back from the Jacs.
11/2/2021
no instruments runs fine, but run defs with instruments never starts
jas»initWindow opens the jas browser
Found error in jacs.log. Nothing wrong with sim startup code.
Loaded JwrGdkServerExtensions app 162.33
JwvSimAuditInfo creates tables like J20211102202345343002_1 on the fly.
The AUDIT_OBJECT in JwvSimAuditInfo is a JwvString which blows up because I am writing BLOBs not LONG_RAWS
The write is initiated by the Jacs before the sim starts so I cannot debug.
Solution 1 is to ignore writing JwvSimAuditInfo at all and see what else happens.
Solution 2 is to hard code the table create sql which also contains several columns of NUMBER(38).
SQL> desc jwars.J20211102202345343002_1;
Name Null? Type
----------------------------------------- -------- ----------------------------
AUDIT_OBJECT LONG RAW
BEGIN_RECORD_TIME NUMBER
END_RECORD_TIME NUMBER
INSTRUMENT_CATEGORY VARCHAR2(255)
INSTRUMENT_COUNT NUMBER(38)
INSTRUMENT_NAME VARCHAR2(255)
LONG_RAW_SIZE NUMBER(38)
REPLICATION_ID NOT NULL NUMBER(38)
RUN_ID NOT NULL CHAR(20)
SEQUENCER_ID NOT NULL NUMBER(38)
SIMULATION_TIME VARCHAR2(20)
11/1/2021
changed number(38) four times in jwars_simrepinfo_history
changed number(38) three times in chemical_clouds_table
changed preflevel in jwars_preferences from number(38) to number
changed report_id in jwars_reportdefinitions from number(38) to number
changed long_raw_size in scenario from number(38) to number
changed long_raw_size in excursion_scenario from number(38) to number
changed long_raw_size in excursion_scenario_component from number(38) to number
descJwarsTables.bat finds 0 occurrences of number(38)
in vast 7.5, the oracle structure was this:
AbtOracleAbstractDatabaseConnection
AbtOracle8DatabaseConnection
JwOracle10DatabaseConnection
AbtOracleDatabaseConnection
JwOracleDatabaseConnection
AbtOracleAbstractDatabaseConnection
AbtOracle8DatabaseConnection
JwOracle10DatabaseConnection
AbtOracleDatabaseConnection
Using JwOracleDatabaseConnection cause classes such as JwAbtRow and JwCursor to be used, perhaps just to handle long raws.
Copied 8 methods from AbtOracleDatabaqseConnection in app JwrAbtOracleExtensions to AbtOracle10DatabaqseConnection .
Plan to convert blobs
10/22/2021
The blobs in env_scenario_grid were probably created on a Unix machine.
7.5 ObjectDumper dumps cannot be read by Vast 10. Big gpf.
7.5 ObjectDumper dumps that are read by 7.5 and then dumped again are different and can be read by Vast 10.
6/9/2021
getting gpf in swapper trying to create JwDataObject environment from scenario_grid blob in env_scenario_grid row
get gpf with buffersize of 1,000,100 and 10,000.
dumped the blob byteArray from the sim and from a script. Both are the same.
reloaded env_scenario_grid from original dump file with long raw after truncating the table
the table description said that scenario_grid was BLOB so the dump file converted them during the import
new BLOBs are the same.
reloaded env_scenario_grid from original dump file with long raw after dropping the table
env tablespace was 12 gigs. dropped it and recreated it as 20 megs.
loaded env tables from 11 dump, then ran fixblob on all but env_scenario_grid
use script to grab long raw blob and dump it. it is the same as the previous three
did this again and the LONG RAW blob (size 15824) was the same, so bufferSize on lobLocator does not matter
ObjectLoader was changed from 7.5
loaded 7.5 ObjectLoader version and still got gpf
need test case
6/8/2021
env_scenario_grid still has a long raw column
there are 7 layer_heights
there are 3 lvl_nm
there are 2 scenario_nms
there are 50 param_nms
system table all_tab_columns has info about all table
-- find all jas tables with a long raw
* select table_name from all_tab_columns
* where owner = 'JWARS' and data_type = 'LONG RAW';
i misread my tbusage script which showed tablespace usage valuse of 33% as free
added 10 gigs to temp tablespace. no joy,
added extra datafile to env tablespace and to_blob ran fine.
no way that oracle conversion failed.
must be a read error.
6/7/2021
getting ora-24374 only on env table
ORA-24374: define not done before fetch or execute and fetch
ran dbv on env dbf and no problems found
C:\app\Jas\oradata\Jas>dbv file=jas_environment001.dbf userid=system\jwars logfile=env.log feedback=1
6/6/2021
fix namePrettyPrint
put jasexp.icx and jass.ini in docs
libmi.dll not found with full path name (C:\Users\Jas\Documents\Instantiations\VA Smalltalk\dev92Jso).
try with dos name (no spaces unquoted)
found when put in temp.
getting ORA-24374 when trying to read jwars.ENV_SCENARIO_GRID
call is from AbtDatabaseConnection»resultTableFromQuerySpec: aQuerySpec withValues: aValueCollection ifError: anErrorBlock
"values"
('SEA(15S-15N,90E-150E)'
'1JAN97-30MAY98'
'Illumination'
'Surface of Earth and Sea'
0.0
361)
"initial sql"
'SELECT * FROM jwars.ENV_SCENARIO_GRID WHERE AREA_NM= :AREA_NM
AND SCENARIO_NM= :SCENARIO_NM
AND PARAM_NM= :PARAM_NM
AND LVL_NM= :LVL_NM
AND LAYER_HEIGHT= :LAYER_HEIGHT
AND TIME_INDEX= :TIME_INDEX'
"result sql"
SELECT * FROM jwars.ENV_SCENARIO_GRID
WHERE AREA_NM= 'SEA(15S-15N,90E-150E)'
AND SCENARIO_NM= '1JAN97-30MAY98'
AND PARAM_NM= 'Illumination'
AND LVL_NM= 'Surface of Earth and Sea'
AND LAYER_HEIGHT= 0.0
AND TIME_INDEX= 361
6/4/2021
number(38) is causing problems. solution seems to be alter table modift (number38Column number)
in jwars_simrepInfo, altered long_raw_size and replication_id
in jwars_simruninfo, altered run_replications
in jwars_simruninfo_history, altered run_replications and sim_duration_days
put breakpoint in prepareForSimulation
remove mod from String»asNumber in VwCompatibility
6/3/2021
jasexp.icx blow up running a scenario when trying to write a construct a row for jwars_simruninfo.
run_replications is decimal with about 38 digits of precision that is seen as ScaledDecimal and is too big.
it is seen by oracle 10 and vast 7.5 as abtPackedDecimal and works fine.
the hack was to modify String»asNumber in VwCompatibility to take just the first 24 digits.
could not add a row to jwars_simruninfo because the index was toasted
rebuilt the index and all was well.
5/31/2021
5/27/2021
ran fixBlob script and most tables ok.
removed refs to selectUniqueSql
latest version of allocateAndPrepareOSMemoryForParameters:ifError: blows up
use 7/23/2019 version
instrument browser dnu indexCountForTable:
message categories in run defs shows nothing. ←– message categories are OK
reloaded db from dump file
ran fixblobs on all but env_scenario_grid (364599 rows 4.7Gs) and tl_location_to_track_grid (535 rows 653M)
WMD HPAC blow ups with dnu
SCL Data blows up with no data
5/26/2021
found two errors in ociLobRead*
first was not testing for oci_Needs_data and oci_success, so only the first 1000 bytes were read
second was the memcopy started at 1 instead of 0, producing bad checksum errors when blobs > 1000 were successfully read.
I fixed the first and bob fixed the second.
I reloaded all jwars_preferences from the object dump file into BLOBs and can now read all 47 objects.
checksum errors were killing me after to_lob conversions. time to try again.
5/21/2021
5/20/2021
modified select to use FOR UPDATE
wrote 47 jwars_preferences objects using exp image
get bad checksum reading them with dev image
wrote them again with ObjectDumper instead of JwObjectDumper.
read them back with ObjectLoader instead of JwObjectLoader and still got checksum error.
the object dump file that is input was written by OjbectDumper.
there are only two jwars_preferences objects bigger than 1000 bytes.
the 'color' 4.0 one is 3472 and is failing on the load.
tried increasing lobWrite buffer size from 1000 to 10000 but no joy.
5/18/2021
5/17/2021
5/16/2021
5/14/2021
do the runsetup_comp example on xp using vast 75 and oracle 10
do it on xp using vast75 and oracle 11 and db8 then db10
then do it on win10 using vast10 and oracle 11 and db8 then db10
5/11/2021
used devExp image to dump an OrderedCollection of jwars_preferences into a file.
then recreated the table with pref_object as BLOB
read the objects back in and did each insert which failed with ORA-22275 invalid lob locator
5/10/2021
created abc table with a blob column and populated it with 14 rows from scenario_component;
doing a select * from abc in a workspace blows up in AbtOracleBlobField»descriptor dnu allocateDescriptor: because AbtSystem activeDatabaseConnection is nil.
5/9/2021
JwOracleDatabaseManager is referenced in dbmClass and only if reuseCursor is true.
setting reuseCursor to false blows up because AbtOracleDatabaseManager cannot find Oracle7 dll.
setting reuseCursor to true uses AbtOracle10* but blows up with ORA-24375 using big blob.
AbtOracle8DatabaseManager and JwOracleDatabaseManager both do the same
reuseCursor does a lot of tricky stuff so we will set it to false and my remove it.
JwrOracleSingleDatabaseConnectionManager is the manager in 7 methods in 7.5 and ok in 10.
No references to AbtOracle8DatabaseManager in 7.5 or 10.
This gets ORA-00997: illegal use of LONG datatype
CREATE TABLE ABC (
SCENARIO_COMPONENT LONG RAW,
SCENARIO_NAME VARCHAR2(255) NOT NULL,
DATE_TIME CHAR(20) NOT NULL,
COMPONENT_TYPE VARCHAR2(255) NOT NULL,
CREATED_BY VARCHAR2(30) NOT NULL,
LONG_RAW_SIZE NUMBER NOT NULL,
ID VARCHAR2(500)
);
So perhaps the imp of Oracle 10 data with LongRaws into Oracle 11 is successfull because imp is dumb, but longRaws are indeed obsolete in Oracle 11.
used the following to copy jwarsPrefs to a new table with a blob instead.
deleted the old prefs table and index and renamed the new.
still get bad checksum on object loader.
CREATE TABLE JWARS_PreferencesNew (
Pref_Object BLOB,
Owner VarChar(20) not null,
Window VarChar(50) not null,
Category VarChar(20) not null,
PrefLevel INTEGER,
CONSTRAINT preference_pkeyNEW PRIMARY KEY (Owner, Window, Category)
USING INDEX TABLESPACE jwars_index
STORAGE (INITIAL 50K
NEXT 50K
MINEXTENTS 1
MAXEXTENTS 121
PCTINCREASE 0))
TABLESPACE jwars_primary
STORAGE (INITIAL 100K
NEXT 100K
MINEXTENTS 1
MAXEXTENTS 512
PCTINCREASE 0);
REM -- Grant all privileges to the specified role for this table.
GRANT ALL ON JWARS_PreferencesNew TO jwarsdb_role;
REM -- Add comments about the table.
COMMENT ON TABLE JWARS_PreferencesNew IS
'Holds information about preferences to be used by browsers.';
insert into JWARS_PreferencesNew select to_lob(Pref_Object),
Owner,
Window,
Category,
PrefLevel from JWARS_Preferences;
select count(*) from JWARS_PreferencesNew;
5/4/2021
hci packaged on ayeone vm still fails
installed vast 75 on xp, packaged with new 3.21, hci still failed.
commented out reference to JwVJads in JwSimulationSystem»runHeadless
packager did not pick up scratch edition
It did pick up versioned edition of JwSimulationSystem and still failed.
packager produces different size icx files and sometimes does not pick up config map name for version
The above relates to packaging hci on xp 7.5
AbtOracleDatabaseManager is still the dbmClass of GdkOracleDatabaseConnectionManager in 9.2
but I change validateOraclePassword:* to use AbtOracle10DatabaseManager.
It was previously AbtOracleDatabaseManager
jas dev jso 102.c is when i rearranged the oracle hierarchy.
AbtDatabaseManager class»libraryName is wher the oracle dll is referenced.
there is an oracle 7 and an oracle 8 library name but there is no 10, so the dll name is probably the same
both oracle 8 and 10 db manager have a displayName method
8 says Oracle 8 and 10 says Oracle 10g/11g
ayeone vm has oracle 10.2.0.1.0
5/3/2021
found va 75 on lenovo
it is not an upgrade and neither is va 752.
delete the reg keys and 75 at least installs.
the whci.icx packaged on levovo failed in the same way.
does the packager look at the loaded apps or load the app in the config map?
#2 seem to be unlikely.
4/29/2021
4/28/2021
packaged hci has older verion of launch: with just two temps.
.es files show that the correct version of JwvSimulationStstem is loaded.
tried this with a clean stock_icx, load map then package
tried putting self halt in launch: method.
tried with clean abt.icx.
looks like I have never successfully packaged HCI
4/27/2021
added code to JwgSimulationMonitor to print the contents of the blue run window out to simMonitor.log when the sim finishes.
made launch: print out more info
repackaged hci with new launch code.
hci version is entire config map name
in run setup, if you uncheck the “use hci version for sim” box,you must pick a version from the dropdown list or the transform check will blow up.
4/26/2021
consider adding isLocal to jacs.ini and forcing the -j parameter to be -j127.0.0.1 if it is true.
This would solve the network connection problem.
reportSimConfigsInDirectory:* is where jacs looks for wh.sim*.icx it can run.
latest jacs map looks at isLocal in jacs.ini and then uses 127.0.0.1 for jacs and env address when starting sims.
sims fail because packaged image still has old launch: code which does not work with hostnames like e6420-xp.hsd1.wv.comcast.net
a possible repackaged hci.icx would solve this problem
4/25/2021
added several logDebug: statements to code that kicks off the sim
kick off will fail if no env server found. needs decent error message.
fixed launch in open image so it will handle a network connection
packaged hci fail to launch when network connection is active
it gets e6420-xp.hsd1.wv.comcast.net and 'Local' which resolves to 10.0.0.252 instead of 127.0.0.1 but still fails.
4/24/2021
gdk.cat should be in jas_oi or in nls
AbtDatabaseManager»startTraceOn: aFileName will log db calls
AbtOracleMultiRowCursor, AbtOracleDatabaseConnection and AbtOracleDatabaseManager are being used.
No JwCursor or JwrDatabaseConnection.
AbtOracleDatabaseManager is the dbmClass of GdkOracleDatabaseConnectionManager
AbtOracleAbstractDatabaseManager subclass: #AbtOracleDatabaseManager
AbtOracleAbstractDatabaseManager subclass: #AbtOracle8DatabaseManager
AbtOracle8DatabaseManager subclass: #AbtOracle10DatabaseManager
GdkOracleSingleDatabaseConnectionManager subclass: #JwrOracleSingleDatabaseConnectionManager
JwrOracleSingleDatabaseConnectionManager is not in the AbtOracleAbstractDatabaseManager hierarchy and has no database connection
Its dbmCkass is JwOracleDatabaseManager is reuseCursor is true.
JwOracleDatabaseManager does fancy stuff with JwAbtOracleRow
reuseCursor should be removed.
04/17/21
tried running ora2sql32 on chuck's machine and got no scenario records.
truncated all 5 scenario* tables on Win10 Oracle11.
used dumpChuckVista(bat,sql) to dump scenario* tables from Chuck;s Vista machine.
Export: Release 10.2.0.3.0 -0 Production on Saturday, 17 April, 2021 14:22:08
Copyright (c) 2003, 2005, Oracle. All rights reserved.
Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 -0 Production
With the Partitioning, OLAP and Data Mining options
Starting"JWARS.""SYS_EXPORT_TABLE_01"":"jwars/********@jas directory=xyzzy
tables=(COMM_MSG_TYPE,COMM_MSG_TYPE_ASSOC,DESCRIPTOR_ASSOCIATIONS,
DESCRIPTOR_CLASSIFICATION,DESCRIPTOR_DEFINITION,DESCRIPTOR_RELATIONSHIPS,
JWARS_PREFERENCES,JWARS_REPORTDEFINITIONS,JWARS_SIMREPINFO,JWARS_SIMRUNINFO,
RUN_SETUP_COMPONENTS,SCENARIO,SCENARIO_COMPONENT,SCENARIO_COMPONENT_LOG,
SCENARIO_COMPONENT_REF,SCENARIO_LOG) dumpfile=chuckVista.dmp logfile=chuckVista.log
exported "JWARS.""COMM_MSG_TYPE_ASSOC""" 7.242KB 47rows
exported "JWARS.""COMM_MSG_TYPE""" 37.02 KB 105 rows
exported "JWARS.""DESCRIPTOR_ASSOCIATIONS""" 229.4 KB 7804 rows
exported "JWARS.""DESCRIPTOR_CLASSIFICATION""" 0 KB 0 rows
exported "JWARS.""DESCRIPTOR_DEFINITION""" 302.7 KB 1197 rows
exported "JWARS.""DESCRIPTOR_RELATIONSHIPS""" 5.656 KB 22 rows
exported "JWARS.""JWARS_PREFERENCES""" 158.2 KB 42 rows
exported "JWARS.""JWARS_REPORTDEFINITIONS""" 471.3 KB 119 rows
exported "JWARS.""JWARS_SIMREPINFO""" 970.5 KB 80 rows
exported "JWARS.""JWARS_SIMRUNINFO""" 9.607 MB 125 rows
exported "JWARS.""RUN_SETUP_COMPONENTS""" 1.645 MB 223 rows
exported "JWARS.""SCENARIO_COMPONENT_LOG""" 24.37 KB 312 rows
exported "JWARS.""SCENARIO_COMPONENT_REF""" 1.16 MB 6033 rows
exported "JWARS.""SCENARIO_COMPONENT""" 101 MB 2955 rows
exported "JWARS.""SCENARIO_LOG""" 23.42 KB 100 rows
exported "JWARS.""SCENARIO""" 241.6 KB 34 rows
Job "JWARS.""SYS_EXPORT_TABLE_01""" successfully completed at 14:22:22
These are the MS table counts on Win10 after db restore using jwars.sql
jwars.sql is a full database dump using direct ora2sql in oracle -> MS transfer and it has a lot of junk in it.
<code>
table_name table_rows
ab_omni_noise 1
ab_shipping_noise_grid 1
ab_spectral_curve 1
ambient_noise_area 1
chem_clouds_munition_agent 0
chemical_clouds_table 0
comm_msg_type 94
comm_msg_type_assoc 47
descriptor_associations 16930
descriptor_classification 1
descriptor_definition 2385
descriptor_relationships 22
env_area 2
env_lvl 4
env_param 0
env_province_grid 1
env_reverberation_area 1
env_scenario 2
env_scenario_grid 364617
env_scenario_param 116
env_scenario_param_layer 217
env_season 4
env_seasonal_grid 7
env_shipping_noise_level 11
env_shipping_noise_level_area 1
env_ss_volume 175
env_static_grid 2
env_static_terrain_grid 466
excursion_scenario 0
excursion_scenario_component 0
geo_area_triangulation 2
geo_area_triangulation_new 9
geo_area_triangulation_ref 2
geo_area_triangulation_ref_new 9
jar_databases 5
jar_table 3183
jar_table_column 13346
jar_table_structure 12911
jarext_table_altered 0
jarext_table_columns 0
jarext_table_component 0
jarext_table_pk 0
jars 4
jwars_components 185
jwars_db_install 1
jwars_preferences 160
jwars_reportdefinitions 171
jwars_simrepinfo 46
jwars_simrepinfo_history 17
jwars_simruninfo 57
jwars_simruninfo_history 24
mob_bse_list 12253
mob_cntry 277
mob_equipment_char 267
mob_force 13260
mob_geo_loc_typ 68
mob_geofile 56676
mob_resupply 16291
mob_sealift_mrs_ships 284
mob_srfdetail 79666
mob_srfforcecat 11476
mob_state 58
mob_truck_classification 101
mob_tuchaf2 9965
mob_tuchaf3 48578
run_setup_components 915
scenario 705
scenario_component 7972
scenario_component_log 2810
scenario_component_ref 34394
scenario_log 1474
tl_area 1
tl_loc_to_track_time_index 12
tl_location_to_track_grid 592
tl_location_to_track_id_grid 0
tl_params_to_tl_time_index 12
tl_scenario 2
tl_track_to_loss_curve 12
tranmission_loss 105
us_jwars_lins 0
04/16/21
Most of the table counts in MS Win10 are ok with the exception of scenario_component which has just 757 of 3600, because ora2sql quits after an Ora-01406 error.
Tried to explore about 10 scenarios and they all blew up with missing Force data.
This means that the data from Win10 Oracle11 is useless unless Oleg fixes Ora01046.
The import into MS of tables such as descriptor_definition, jwars_preferences, and run_setup_components changed a double column to a bigInt and Jas blows up with SmallInteger does not understand abrAsArray.
This seemed to happened in the Oracle→MS transfer because I chose table overwrite.
The solution would be to find all tables with bigInts and change them to doubles.
alter table someTable modify column columnNamed newType (doouble, etc.)
Then use ora2sql to dump the data to sql files instead of doing a straight transfer.
Then import them into the existing MS tables after removing any create table sql from the exported files OR changing the table creation to the right data type.
Use ora2sql to dump individual files from Chuck's machine after doing a count.
Used mysql -uroot -pjwars jwars < jwars.sql to restore db from mysql vm
run ms table description again and compared.
the complete reload of jwars.sql used double instead of bigint, and a latin swedish charset instead of utf-8.
it also used longBlob instead of mediumBlob which is restricted to 16M.
04/15/21
From the comment on Jas DEV: MYSQL R2.0 SR2.53 1.5 (12/8/2010 1:07:07 PM): changed AbtOdbcDatabaseConnection»executeWithParmsQuerySpec: withValues: ifError: to not use a cached cursor if doing an INSERT. This fixes transforms, cloning runDefs, and writing instrument data.
Fix error message in GdkOdbcMySqlDatabaseConnectionManager»connectWithAlias:showErrors:
use older version of »launch: that does a better job of finding the host name.
jas dev blows up in initializeInfrastructure on a call to the env server.
many problems packaging env.icx
tried repackaging jacs 1.01c which was running ok.
got errors about missing class SstInvocationHandler in a packagingRule and OpenSSLCryptoInterfaceMAC missing some ssl constants also in a packagingRule
also the packagingRules that set the map name, etc. did not complle
finally I loaded Jacs Pks 1.01c in the dev image and then loaded it again in the xd image.
everything was just as before. not sure why.
jas open image still blowing up on (I think) an environmental call.
no-oped the scenario transform in hci and repackaged.
pkg hci blows up reading runSetup object.
maybe needs blobPatch
created a new runSetupObject and then added message categories.
blew up on the rewrite probably because the size was 4716.
maybe look at executeWithParmsQuerySpec: above
04/14/21
aRuleCreationInterface doNotReduceClassNamed: #EsIdentitySet.
versionName and configMapName are in JwSimulationSystem
added do not reduce EsIdentitySet to (finally) the correct app
reverted to original oracleBlobPatch:. runDefs are created with blobs ~300, and then subsequent updates of, messages for instance, fail because the new blob size is 1204.
make initial blob be 4k and all is well.
remove JwSimulationSystem»setOdbcMySqlConnectionmanagerMaxCacheSize
Fix error message in GdkOdbcMySqlDatabaseConnectionManager»connectWithAlias:showErrors:
04/13/21
[Database]
DatabaseType=MySQL
LogonId=jwars
Password=password
ServerName=jwars
Make Jacs pick up all four of these parameters
assignDBLogon should then test that DatabastType = “mySql” or “Oracle”
changed [Odbc] in jas.ini to [MySql]
JwrDatabaseSystem»assignDBLogonSpec calls Jacs»requestDbCommonAccountSpec
in Jacs, removed default DatabaseType of 'Oracle' in JwvJACSDBManagement»databaseServerType.
new jacs.ini
Changed GdkOdbcMySqlDatabaseConnectionManager»databaseHeading from 'Odbc' to 'MySql'
Added test of DatabaseType in JwrDatabaseSystem»assignDBLogonSpec
04/12/21
++++++++++++++++++ ODBC
| anAbtQuerySpec resultTable conn logonSpec results stmt |
logonSpec := AbtDatabaseLogonSpec id: 'jwars'
password: 'password' server: 'jwars'.
" the dsn below (jwars) is the name of the ODBC data source
that holds the user/password/db name.
conn := AbtOdbcDatabaseManager new connectToDataSourceNamed: 'jwars'
logonSpec: logonSpec ifError: [:e| self halt].
Changed this in jas.ini:
[Rule Data]
IwedaPath=C:\ProgramData\Jas\client\iweda
created GdkAbtDbmOdbcBaseAppExtensions and JwrAbtOdbcExtensions app
rewrite or preferences works - can change and save Confirm Exit
can browse all Preferences and Categories
Studies was blowing up on saves until I copied AbtOracleRow»data to AbtOdbcRow
Can manipulate Run Defs including clones.
fixed Measures looking at instruments by adding hciSystem to JwvRunSetupDataCollection
EEAs works.
Clone works on Measures and EEAs.
Copied last Jas Packaging: HCI map to Jas Packaging: HCI MySQl and updated apps
04/11/21
ran createUser and createGrant on MS
updated JSO MS map to 9.2
loaded last JSO ODBC map and then compared to newset map.
remove dbmClass from JwrOdbcSingleConnectionManager
latest error is cannot find odbc drivers.
-
-
04/10/21
setting up environments for 7.5 and 9.2
the missing libslac message, in areCriticalTokensValid, should show the whole path.
NO! libslac likes to live above 7.5 , 8,2, etc.
did not seem to work even with FQDN. needs work.
oddly millibary is getting found, yes?
MS VM was running with a jacs that said it was 3.20 but which responded to a jacs.ini with ms settings.
updated jacs to 9.2.
MS needs a user jwars/password to logon.
error in xd map loading picking up an older map.
was updating the last odbc jso map.
finally took last 9.2 jacs map and updated it.
the names appeared to be the same but the timestamps must have been different.
created a new jacs with odbc support.
logon tried to connect to MS but fails because no jwars user
04/09/21
start oracle and the first listener in the services table.
check table counts in A1 Oracle
use expdp to dump all non-static table data only from A1 Oracle
on A1 in dir \mysql\expdp, created dumpA1jas10TableData.bat to create A1JAS10TABLEDATA(1-3).DMP, then copied it to shared dir M6800\temp.
also created loadA1jas10TableData.bat
recreate win10 oracle in \temp\sql and\temp\xyzzy.
use impdp to import the above data into win10 Oracle
win10 Oracle counts are much higher for scenario* tables.
save win10 my.ini
re-install ms562
save ibdata1
use new ora2sql to dump table structs from win10 Oracle then import into ms562
use new ora2sql to dump table data only from win10 oracle and them import into ms562
check table counts in ms562
in a clean ibdata1, did mysql -u -p jwars < jwar.sql
counts are not great but tables are created and it looks like env* and tl* got loaded.
did 6 seperate ora2sql dumps of table data, other than env* and tl*.
then dir mysql -u -p jwars < each.sql for each.
table counts show 800+ scenario rows even though Oracle has just 767.
04/08/21
copied a1 ibdata1, jwars directory, deleted log files and MS 562 restarted on Win10, but could not see any files in jwars.
restored default/installed MS 562 ibdata1, deleted log files, and started MS 562.
did create database jwars in mysql.
then did mysql -uroot -pjwars jwars < jwars.sql which is the dump from yesterday.
got Error: 2006 MySQL Server has Gone Away.
added this to my.ini and restarted server
[mysqld]
max_allowed_packet=67108864
did mysql -uroot -pjwars jwars < jwars.sql twice and got different table counts from A1. not reliable.
the table row counts when sustituting the A1 ibdata1 is less than the first load using mysql
A1 ibdata1 is 15,599,616 and the new ibdata1 after input of jwars.sql is 15,001,600
tried mysql -uroot -pjwars jwars scenario < jwars.sql to load just the scenario table, but the input sql file must have the same name as the table.
there seems to be no -tables=table1, table2, … switch
mysql -uroot -pjwars jwars scenario < scenario.sql does not seem to work.
need to use mysqlimport -uroot -pjwars jwars scenario < scenario.sql.
all of the current files do a drop table and then a table create.
ran mysqldump -u -p –no-create-info jwars scenario > scenario.sql on MS VM.
04/07/21
did something like this for each table
* mysqldump -u... -p... jwars --hex-blob tableName > tableName.sql
moved both dump and dumpHex from ayeone vm to Win10 mysql dir.
downloaded and installed latest MySql community edition (8.0.?) which is 64 bit.
took developer option and then all install defaults but did not create the (optional) router
added C:\Program Files\MySQL\MySQL Server 8.0\bin to win10 path so I can use the command line.
createdatabase, createUser, and createGrant all worked
the loading of table data one by one in loadDatabase failed because Error: 1290, The MySQL server is running with the –secure-file-priv option which means it will only do imports from a specific directory.
Find the directory thus:
mysql> show variables like "secure_file_priv";
+------------------+------------------------------------------------+
| Variable_name | Value |
+------------------+------------------------------------------------+
| secure_file_priv | C:\ProgramData\MySQL\MySQL Server 8.0\Uploads\ |
+------------------+------------------------------------------------+
1 row in set (0.00 sec)
edited C:\ProgramData\MySQL\MySQL Server 8.0\myi.in and changed this:
# Secure File Priv.
secure-file-priv="C:/ProgramData/MySQL/MySQL Server 8.0/Uploads"
to this:
# Secure File Priv.
secure-file-priv=""
and restarted the MySql server service.
getting many wacky errors (1261, 1406,1265) doing mysqlimport to ms 8.0 using 5.1 exported files.
the table creation sql was created 5/2009 by a 2006 version of ora2sql. it needed some hand jamming.
downloaded and installed MS community edition (CE) 5.5 which is also 64 bit.
stopped the 5.5 MS server, copied the jwars directory (with the .frm files) into its data directory, replaced the installed 5.5 ibdata1 with the jwars one, deleted all the log files and started the MS server.
Went into MS 8 Workebench and added a connection to the 5.5 MS server running on 3307.
Hooked 5.5 and 8.0 up and tried to do the import wizard from 5.5 to 8.0 which failed.
Used the MS 8 Workbench to do a jwars database dump.
deleted both MS 5.5 and 8.0
Found the MS 5.7.2 was the last 32 bit version, but no MSI installer.
Found MS 5.6.2 was the last 32 bit version with an installer.
Downloaded and installed 5.6.2. Copied my ibdata1 file as above and MS 5.6.2 server started
Bought latest version of Ora2Sql which was 64 bit and would not install on A1 XP VM.
Got ora2sql32 from tech support
04/06/21
mysqldump -u… -p… jwars > jwars.sql
this ^^ tries to dump the entire jwars database and keeps dying with a 2013 error.
do something like this for each table
* mysqldump -u... -p... jwars tableName > tableName.sql
for A in (jacs jads jas) do command /c type A.ini
copied all 16 gigs of ayeone mysql to win10.
no ibdata1 file or .frm files.
04/05/21
The Ayeone VM has an Oracle Jas10 database with 500+ scenarios.
There are batch files and sql in c:\jas\sql.
Dumped the entire schema to dumpAyeoneJas10JwarsSchema.dmp.
Dumped the 6 scenario* tables to dumpAyeoneJas10ScenarioTables.dmp.
mysql> select long_raw_size from scenario where scenario_name = "ta SEA VIKING R
2.0 SR3.20 TL ATP 03";
+---------------+
| long_raw_size |
+---------------+
| 6776 |
+---------------+
1 row in set (0.00 sec)
mysql> select long_raw_size from scenario where scenario_name = "ta SEA VIKING R
2.0 SR3.20 TL ATP 03";
+---------------+
| long_raw_size |
+---------------+
| 6776 |
+---------------+
1 row in set (0.00 sec)
mysql> select LONG_RAW_SIZE from scenario where scenario_name = "ta SEA VIKING R
2.0 SR3.20 TL ATP 03";
+---------------+
| LONG_RAW_SIZE |
+---------------+
| 6776 |
+---------------+
1 row in set (0.00 sec)
mysql> select LONG_RAW_SIZE from SCENARIO where scenario_name = "ta SEA VIKING R
2.0 SR3.20 TL ATP 03";
+---------------+
| LONG_RAW_SIZE |
+---------------+
| 6776 |
+---------------+
1 row in set (0.00 sec)
mysql> select LONG_RAW_SIZE from SCENARIO where SCENARIO_NAME = "ta SEA VIKING R
2.0 SR3.20 TL ATP 03";
+---------------+
| LONG_RAW_SIZE |
+---------------+
| 6776 |
+---------------+
1 row in set (0.00 sec)
04/04/21
Back to MySQL
mysql -uroot -pjwars
use jwars;
show all tables;
show databases; "databases are the same as schemas"
show schemas; "same as above"
SELECT table_name, table_rows FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_SCHEMA = 'jwars';
From mySQL VM, dumped last 3 Jas Dev MySql maps and DMM SUnit MySwl test.
looking at 2014 latest unreleased app, all look ok except for the first one.
It looks like I fixed fullTableName to nor prepend 'user_' when user = jwars
Also looks like launch is fixed to ignore wifi hostname
Created and exported new Jas Dev MySql map.
Found a duplicate of this map on drive e:. Need to add some identifying info.
04/03/21
Sent this to Chuck and Paul. JasANdOracle.doc.
JAS and Oracle
At the end of 2010, JAS (hereinafter JAS8) was running on Windows XP (WinXP) using Visual Age Smalltalk 7.5 (VAST7) and talking to an Oracle 10 (Oracle10) database. This document describes my efforts to making JAS run (as JAS10) on Windows 10 (Win10) using Visual Age Smalltalk 9.2 (VAST9) and talking to an Oracle 11 (Oracle11) database.
Oracle8 was released in 1998 and Oracle10 was released in 2004. Since then, there have been Oracle versions 11, 12, 18, and 19. Oracle 11 is the last one I could find that has a 32 bit version. JAS is a 32 bit application and can “talk” only to a 32 bit database.
Smalltalk and Oracle are fundamentally different. Smalltalk has objects which usually are composed of other objects. A Person object might have a Name object and an Address objects and so on. Oracle has tables with typed (string, integer, etc.) columns called Name, Address, and so on. To store Smalltalk objects in Oracle, it is necessary to “map” the parts of a Smalltalk object to columns in an Oracle table or tables.
The problem for the designers of these mappings, Ron Painter, Kuong Lai, Dave Cedel, Chris Wolcott, and John Sheffler) is that there are hundreds, perhaps thousands of different Smalltalk objects in JAS - BSEs, their Assets, etc - and it would require hundreds of Oracle tables to store them all. So they decided to use BLOBs.
BLOB stands for binary large object. A BLOB is simply a series of bytes which you can store into and retrieve from Oracle. For instance, you could store the bytes of a JPEG photograph in a BLOB. Oracle knows nothing about what the BLOB represents. It is completely up to the application program to convert the BLOB data into something useful. As I said previously, Oracle columns are typed. The type in Oracle8 for a BLOB is a LONG_RAW.
Consider the example of user preferences in JAS. There are 47 of them, and they record the user's desired settings for things such display order, window layout, etc. To get around the necessity of having 47 different Oracle tables, one for each of the 47 Smalltalk objects, the designers came up with the idea of a single table that could stores all of the 47 preferences by storing them in LONG_RAW columns.
A simplified description of the table would be:
CREATE TABLE JAS_PREFERENCES (
KEY INTEGER
DATA LONG_RAW)
So to store the Smalltalk object that represents the desired window layout, say WindowLayout, as the first preference, you would set the JAS_PREFERENCES key column to 1, and use the Smalltalk ObjectDumper to dump the bytes that represent the WindowLayout object into the DATA column, and then write the row to the JAS_PREFERENCES table.
To read the WindowLayout object from Oracle , you would use SQL like this to get the desired row from the JAS_PREFERENCES table:
SELECT * FROM JAS_PREFERENCES WHERE KEY = 1;
and then use the Smalltalk ObjectLoader to load the bytes in the DATA column and use them to reconstruct the Smalltalk WindowLayout object.
JAS8 and VAST7 with Oracle10 worked quite well for a long time. Then came Win7.
Oracle10 will not install on Win7, period. So I used the Oracle8 dump utility to dump the entire JAS8 database from Chuck's Vista machine, installed Oracle11, and used its import utility to rebuild the database. I updated all the JAS8 code that was in VAST7 to VAST9. The VAST7 code used Oracle8 functions. The VAST9 codes uses Oracle10 functions which are a superset of the Oracle8 functions. Meanwhile Oracle deprecated the use of the LONG_RAW column strongly suggesting that you convert such columns to their new BLOB data type.
Technical discussion follows. Before calling Oracle to fetch a row of data using SQL such as the above, you must first define/allocate some space in your program where Oracle can put the retrieved data. The VAST7 and VAST9 code looks at the table columns and tries to allocate some space for each column. This is easy for simple Oracle types such as INTEGER, because Oracle knows how big an INTEGER is. It is not so simple for a LONG_RAW column, because Oracle has no idea how big the data in the LONG_RAW column is.
JAS has some large LONG_RAW columns. The ones in the TRANSMISSION_LOSS table are 29360128 bytes long. The JAS8 code that uses the VAST8 oracle functions forcibly allocates/defines this much memory for each and every LONG_RAW column no matter how big it really is. The problem in JAS9 using VAST10 Oracle functions talking to an Oracle11 database is that the forcible define/allocation of a LONG_RAW of 29360128 bytes fails with a fatal error of ORA-01062: unable to allocate memory for define buffer.
However, removing the requirement that LONG_RAW columns always allocate 29360128 bytes allows tables with smaller LONG_RAWs, such as JAS_PREFERENCES, to load. But all the rows in tables with larger LONG_RAW columns, such as Scenario and Scenario_Component, fail to load at all with error ORA-01406: fetched column value was truncated, because there was not enough space allocated to hold the data in the LONG_RAW column that Oracle fetched.
The VAST code that does this define/allocate is different between VAST7 and VAST9. VAST7 uses Oracle8 functions and VAST9 uses Oracel10 functions. Possible explanation for this failure are:
1. Oracle11 does not allow large defines on LONG_RAW columns or has deprecated their functionality in Oracle11.
2. The VAST9 code using the Oracle10 functions is wrong.
I seriously doubt that the Oracle code is wrong, and I found another error in the Oracle10 functions in VAST9. Perhaps no other VAST9 user deals with large LONG_RAWs.
Oracle has a function called to_lob that converts data in LONG_RAW columns to the newer BLOB data type. I ran this function on the 25 JAS tables that have LONG_RAW columns and there were no Oracle errors in the conversion.
The first Oracle table JAS reads when it starts up is the preferences table mentioned above. JAS never even got the first row loaded because the VAST9 ObjectLoader blew up loading the (now) BLOB column with a bad checksum error, leading me to believe that the to_lob function had done something to the data in the LONG_RAW during the conversion. Another possibility is that the data was dumped into the LONG_RAW column by the VAST7 ObjectDumper and that data cannot be read correctly from the BLOB by the VAST9 ObjectLoader. I think this unlikely because Instantiations was fanatical about not breaking things in earlier versions.
Next I reloaded the Oracle11 database to its original state where BLOB data was in LONG_RAW columns. Read in all the prefences data, created the corresponding Smalltalk objects, then used the VAST9 Smalltalk ObjectDumper to write/dump the objects to a file. Used the Oracle to_lob function to convert the preferences BLOB column from LONG_RAW to BLOB. Used the VAST9 Smalltalk ObjectLoader to read the dumped preference objects fromthe file, created new Smalltalk preference objects, and tried to update just the BLOB column in the database. Got a new error: Error updating JwvPreferenceSettings: (DEFAULT Database Table Viewer displayGroupBox) - AbtError: rc=-1 for 22275 in an AbtOracle10DatabaseConnection at (4/1/2021 9:11:21 PM) ORA-22275: invalid LOB locator specified.
Finally I dropped the preferences table and then recreated it using BLOB as the data type instead of LONG_RAW. The JwvPreferenceSettings class has a method, addAllDefaultPreferences, that creates the default preferences. I executed JwvPreferenceSettings class»addAllDefaultPreferences and got the same 22275 error as above. This leads me to believe that there is something in the original JAS code that talks to Oracle that has problems with VAST9 Oracle10 functions. The updates from VAST7 to VAST9 left this code largely untouched.
I think this last problem could be solved by John Sheffler. However, even if he fixed that, there is still the original problem with the VAST9 Oracle10 function failing when trying to define/allocate 29 megs of memory for a LONG_RAW column fetch. This probably requires help from Instantiations.
04/01/21
used oralce8DbConnection and still got define error.
read all of jwars_preferences into an orderedCollection.
used objectDumper to dump the entire orderedCollection to a file as a single object.
truncated the table and then did a to_lob on the pref_object.
used objectLoader and successfully loaded the orderedCollection of jwars_preference objects.
then spun through the collection and did an insert on each.
the first one failed with a scaled decimal error trying to unpack prefLevel which is an abtOracle8DecimalField into a sql field defined as number(38).
Did a table recreate using new sql which uses integer instead of number(38) and got the same error
desc jwars_preferences says that prefLevel is still number(38)
tried again using objectLoader and successfully loaded the orderedCollection of jwars_preference objects.
then spun through the collection and did a self update: #(value) on each which evidently did update the blob because I still got checksum errors.
this ^^^ is wrong. stepping through the load: method in the debugger we get 'Error updating JwvPreferenceSettings: (DEFAULT Database Table Viewer displayGroupBox) - AbtError: rc=-1 for 22275 in an AbtOracle10DatabaseConnection at (4/1/2021 9:11:21 PM) ORA-22275: invalid LOB locator specified' which means that the update failed but our load: method did not check for any errors.
posted my problem on stackoverflow
04/02/21
in load: method, retrieved each jwars_preferences by key first, updated its value from the loader object's vlaue, and then did update: #(value).
still got ORA-22275 as below.
load: class
| loader objs stream |
(stream := CfsReadFileStream open: (class tableName , '.dmp')) isCfsError
ifTrue: [self error: stream printString].
stream isBytes: true.
loader := ObjectLoader new.
objs := loader loadFromStream: stream.
stream close.
loader hasErrorOccurred ifTrue: [self error: loader currentErrorString].
objs do: [:each | | existingObj |
existingObj := each class retrieveByKey: each.
existingObj value: each value.
existingObj update: #(value)]
changing Number(38) to just NUMBER gets us past the scaled decimal error
running #JwvPreferenceSettings class»addAllDefaultPreferences fails with ORA-22275 which probably means the original Jas Oracle code is at fault somewhere.
03/31/21
after db recreation, jwars_preferences still failing with bad checksum.
doing desc jwars_preferences shows that pref_object i still the converte Blob not the original Long. why?
screwed up and used sql creation scripts from 2017 where I had replaced Longs with Blobs.
ran droptb.sql and then createtb.sql to recreate the tablespaces.
ran impdp system/jwars@jas table_exists_action=append schemas=JWARS directory=xyzzy dumpfile=expdatOracle11LongRaw.dmp logfile=importExpdatOracle11LongRaw.log
this created the tables and indexes with no errors. no need to run any create*sqls.
both jwars_preferences and scenario_components are back to long_raw.
we still get ora-24734 on jwars_preferences when MaximumBlobSize = 29360128, even using all 3 versions of ociStmtFetch2:error:
comment out MaximumBlobSize in jas.ini
jas open now gets to the explorer which means it loads jwars_preferences ok.
now it blows up with an ora-01406 when trying to load Acoustic Sensing in the first AJH scenario which is where we were 3/26.
in jasDevJsoExp, we can fetch all scenario, jwars_preferences, and simRepInfo are all ok with Long_raws < 20k.
No scenarioComponents work.
AbyOracle8Cursor»describeRowIfError: is where the AbtCompoundType is assembled.
AbtOracle10DatabaseConnection»defineWithParam:ifError: fails when trying to define 29 meg long raw with ORA-01062: unable to allocate memory for define buffer.
what are platform limits?
03/30/21
after removing maxBlobSize from jas.ini, all of jwars_preferences loaded ok before conversion to Blobs.
Acoustic Sensing in Characteristics and Performance, which is almost certainly JwvScenarioComponent objects, failed on ora-01406, which I fixed.
But after conversion to Blobs, about 4 of the 41 jwars_preferences failed with blob checksum errors (26) in the Loader.
this also happened with all blobs in JwvSimRepInfo, JwvScenarioStorage, and JwvScenarioComponent.
This means that using to_lob to convert Long Raw to Blob has toasted the checksum of some blobs less than or equal to 4001 (jwars_preferences) and all blobs > 4001.
created new map AbtBaseOracle V 82.0.0 [453]a with fix to AbtOracle10DatabaseConnection»ociLobRead:amount:piece:
Restoring Dump File
all these files are in z:\xxx
ran droptb.sql and then createtb.sql to recreate the tablespaces.
then cd to database\scripts\system\new and ran sys2-4.bat.
jwars_db_roll, jar_role, and user jwars already exist, because they do not live in the tablespace.
cd to database\scripts\user\jwars\new
run createAll.bat
cd \temp
do mkdir xyzzy
copy expdatOracle11LongRaw.dmp into xyzzy
run xyzzy.bat
cd \temp\xyzzy
run impdp system/jwars@jas table_exists_action=append schemas=JWARS directory=xyzzy dumpfile=expdatOracle11LongRaw.dmp logfile=importExpdatOracle11LongRaw.log
this fails to load some env* tables because of missing keys so run it again
impdp system/jwars@jas table_exists_action=append schemas=JWARS directory=xyzzy dumpfile=expdatOracle11LongRaw.dmp logfile=importExpdatOracle11LongRaw2.log
run countRecsInAllTables.bat and compare counts to export log.
env_ss_grid and env_seasonal_grid numbers were off.
drop both tables and then do impdp system/jwars@jas table_exists_action=append tables=jwars.env_ss_volume directory=xyzzy dumpfile=expdatOracle11LongRaw.dmp logfile=importExpdatOracle11LongRaw3.log
note the use above of jwars.table_name
did not run createallIndexes.sql.
could not find addforeignkeys.sql in user mentioned in recreate db 12/18/2017
03/29/21
did an expdp first \jas\app\admin\jas\dpdump\expdatLongRaw.dmp.
used the to_lob oracle method to convert all long raw to blob. took about 15 minutes.
got same same ora-24374 with maxBlobSize.
fixed ora-01496 by modifying AbtOracle10DatabaseConnection»ociLobRead:amount:piece: by testing for OCI_NEEDS_DATA as well as OCI_SUCCESS, though probably only the first one is needed.
in image jasdevjosExp, did retrieveBySql one blob table at a time. see bloblist.txt
All 34 JwScenario rows failed with loader error 26 - bad checkSum.
Only 4 of 41 preferences failed, none with a bytearray size < 4001. some 4001 failed and some worked.
for scenarios and scenarioComponents, all bytearrays were > 4001 and all failed.
both scenario and scenario component have a long_raw_size column
the value from sqlplus 'select long_raw_size from scenario' matches the fetched bytearray size for each row
long_raw_size is not an scenario instVar and is not used on input.
only references to 4000 in jas dev jso are in oraclePadBlobPatch:
no jasdevjso references to 4001
ToDo: NO! create a 7.5 object with name and long raw of about 60k. dump it. then load into 9.2. do vice versa.
03/26/21
notes from approx 12/6/17 imply that only the creation sql files have been modified.
they do NOT say that I changed LONG RAW to BLOB in the database I imported 12/8.
probably need to change Jas refs to AbtOracleLongRawField to AbtCLOBField
blobSize for AbtOracleLongRawField is set in GdkOracleDatabaseConnectionManager»connectWithAlias: anAlias showErrors: showErrors
we get ora-24734 because oracle cannot allocate MaximumBlobSize of 29360128 when doing fetchNextRow on JwvPreferences
we get ora-24734 even with MaximumBlobSize set as low as 100000 in jas.ini.
we get past fetchNextRow on JwvPreferences when MaximumBlobSize in jas.ini is commented out, but now get ora-01406 when trying to load Acoustic Sensing in the first AJH scenario. Debugging shows that the AS blob is ~119k.
Restoring JwOracleSingleDatabaseConnectionManager and subsequent call sequences like 7.5 made getAllPrefernences blow up whne doing modelFromRow:.
Using AbtOracle8DbMgr (instead of 10) gets unsuccessful blob load when loading the Acoustic Sensing blob above.
03/25/21
JwvDatabaseModelsApp 3.94d has soem stuff probably related to CLOBs. Use 3.15 5.17 for now.
JwvDatabaseSystemApp 1.00b is ok.
Use last JSO version of JwvJacsProxyApp SR3.16 3.17
use last JSO version of JwrDataObjectsApp SR2.57 3.21
Z:\jas\Oracle\exportedDataOracle10g has a dump file from Oracle that perhaps has Long raw converted to blobs.
Z:\jas\Oracle\createJas\Database has new user scripts to create tables with new blob format
Similar files are in Z:\jas\Oracle\xpLaptop\new
probably need JwvDatabaseModelsApp 3.94 which has blobspec in it
use this to change lon raws to blobs
create table mytable (pk number, blob_column long)
-- add lots of rows
alter table mytable modify (blob_column clob);
12/25/17
This starts old JACS
C:\Program Files (x86)\JAS\bin\whJACS.exe" -srvName whJACS -i.\whJacs.icx -ini:vmcfg.ini -lwhJacs.log -srvWorkingDir "C:\Program Files (x86)\JAS\jacs
This does not start new JACS
C:\Progra~2\Instan~1\VASmal~1\9.0x86\abt.exe -srvName whJACS -iC:\Progra~2\JAS\jacs\whJacs.icx -ini:vmcfg.ini -lwhJacs.log -srvWorkingDir C:\Progra~2\JAS\jacs
use jasdev.icx - last save 8/28/17 with jas dev jso 3.20 loaded
In XD, Loaded JAS Packaging: Sim R2.0 SR3.20 (with its required maps).
In XD, Loaded JAS Packaging: Jacs R2.0 SR3.20 (with its required maps).
icx was 6973, icx from 8/27 was 6974 —————————————> WORKS!
need to fix some gdk missing messages that the 8/27 icx does not have
12/25/2017 3:51:37 PM | Info : Encryption interface initialized in FIPS mode
25 Dec 2017 15 51 38 | Warning : The setting {EncryptionEnabled} is missing under the section {Oracle}. A default value of false is being used.
25 Dec 2017 15 51 38 | Info : Message Log Launched on jacs.log
25 Dec 2017 15 51 38 | Trace : Launched server listener on 192.168.1.9:8000 [Hostname is Lenovo]
25 Dec 2017 15 51 38 | Trace : **MISSING GdkMsgLaunched in gdk**
25 Dec 2017 15 51 38 | Info : Reading Configuration File - jacs.ini
25 Dec 2017 15 51 38 | Info : Version Information (Below)
-------------------------------------------------------------------
Configuration Map - JAS Packaging: Jacs
TimeStamp - Unknown
Version - Unknown
Package TimeStamp - 25 December 2017 14:34:48
-------------------------------------------------------------------
25 Dec 2017 15 51 38 | Info : Overlay Information (Below)
-------------------------------------------------------------------
Overlays file [JACS Overlays.xml] does not exist.
-------------------------------------------------------------------
25 Dec 2017 15 51 38 | Info : Simulation scheduling will be controlled by Internal Manager.
25 Dec 2017 15 51 38 | Debug : [Sim Version Heartbeat] Sleep until 0300.
25 Dec 2017 15 51 38 | Debug : [Server Heartbeat] Sleep for 30 seconds.
25 Dec 2017 15 51 38 | Info : Reporting for .\whSIM.R20.SR320.icx
25 Dec 2017 15 51 38 | Info : Command -> whSIM.exe -report -j192.168.1.9:8000 -lsimvm.log -ini:vmcfg.ini -i.\whSIM.R20.SR320.icx -nosplash
25 Dec 2017 15 51 41 | Error : **MISSING GdkMsgCannotConnect in gdk**
25 Dec 2017 15 51 46 | Trace : **MISSING GdkMsgClientConnected in gdk**
25 Dec 2017 15 51 46 | Trace : * [Update] CLIENT HEARTBEAT RATE from @ 192.168.1.9:61358
25 Dec 2017 15 51 46 | Trace : **MISSING GdkMsgClientDisconnected in gdk**
25 Dec 2017 15 51 46 | Info : Completed
25 Dec 2017 15 51 48 | Trace : MISSING GdkMsgClientConnected in gdk
Packaging errors make no difference. In file pkgErrors320.txt
12/24/17
Jas 321 [b378] has 8/22 date, but current jacs was created 8/27
hooked up 378 env image to \jas\envy\mgr75With378.dat from early 9/17
imported jacs packaging 3.17, 3.20, and 3.21a/b/c
3.17 does not load - duplicate OSSslStructure
3.20 does not load - duplicate OSSslStructure
3.21a is missing JwrExtensions (8/22/2017 11:20:11 AM) ts=3680853611
3.21b has JwrExtensions ((8/22/2017 7:38:01 PM) ts=3680883481) loads ok
12/21/17
Jacs packaging instructions after last good JSO (3.17) are crap! Missing apps and required maps.
JAS packaging instructions are in Jwx*Packaging. These should be loaded into the Dev image so they show up in XD.
Need to manually add Oracle 10, Crytpo, SSTInvocation, etc.
Look at reduction errors. Should be none. Ideally.
12/20/17
cannot execute any code in an XD image
get next version of JwvJACS/JwSimulationSystem, write it and ts to a classVar, then version release class, app, and config
see EtApplicationManager»versionClasses: classEditions prompting: aSymbol for how to get next version name
12/19/17
DK Development needs Dk Oracle Back End
Dk Oracle Back End needs Dk Object Sharing
Dk Object Sharing needs Dk Networking
Dk Networking needs DK Kernel
packaging jacs
create XD image with TCP/IP, Oracle, External Calls, Web, Windows and both XMLs
in the XD image, load Sim Win
load JAS Packaging Jacs
Select JacNtHeadless from Instructions in Database
whJacs.icx is 7880; august icx (the current one) is 6974; 7.5 (bad version) is 5237
12/8/17
recreate db
drop tb
recreate tb
run system sql 2-4
run createall.sql in user
do not have to disable constraints
create \temp\xyzzy
create or replace directory xyzzy as '\temp\xyzzy'; ←—– this is remembered
impdp table_exists_action=append schemas=JWARS directory=xyzzy dumpfile=expdatFromExpdp.dmp logfile=impJwars.log
run addforeignkeys.sql in user
^^^^^^^^^^^^^ worked 12/9/1
impdp table_exists_action=append tables=JWARS.RUN_SETUP_COMPONENTS directory=xyzzy dumpfile=expdatFromExpdp.dmp logfile=runSetup.log ←— this will rebuild a truncated table
^^^^^^^^^^^^^ worked 12/13/17
12/6/2017
removing MaximumBlobSize=29360128 fro jas.ini gets past ora-24734
change all LONG_RAW to BLOB
case sensitive jwars password is password
integer datatyep is Number(38)
alter table JWARS_SimRunInfo modify Run_Replications int; WORKS
2017 Dec 8 D MacQueen changed NUMBER(38) to NUMBER(10), change LONG_RAW_SIZE to INTEGER, change LONG_RAW to BLOB
12/5/17
loaded jas dev jso 3.78d in jds2
loaded latest GdkOracleSupport and JwrDatabaseSystem to add Oracle8
failed with ora 24734
<c: int32 'Oracle8LibraryName':
without oracle8 → <c: int32 'Oracle7LibraryName':
get key not found (0) when looking up an oracle type
“Modified Dec 8, 2017 D MacQueen”
wServices messageLog logWarning: ←————- use for logging vm type
12/2/17
cloned finalmgr75.dat to reduce size
added manager90.dat from 394
renamed to jas90.dat
11/29/17
changed JwModel class»getAllRowsByQuerySpec: aQuerySpec withValues: valuesDict ifError: errorBlock
changed GdkDatabaseConnectionManager»connectWithAlias: anAlias showErrors: showErrors
fix JwBlock»initializeFromBlockText: blockString inContextOf: anObject ifCompileError: errorBlock
remove hard coded references to AbtOracleDatabaseManager mostly in dbmClass. change to JwvDatabaseSystemApp dbmClass which is AbtOracle8DatabaseManager
comment out auditSuccessfulLogin
comment out auditSimulationExecution
comment out acceptLogonBanner
fix resetDefaultIconManagers - hierarchy change
fix EffectsServerManager»parseParameterFile: aFileStream
AbtOracle8DatabaseManager
<c: int32 'Oracle8LibraryName':OCIStmtExecute pointer pointer pointer uint16 uint16 pointer pointer uint16>
<c: int32 'Oracle8LibraryName':OCIStmtExecute pointer pointer pointer uint16 uint16 pointer pointer uint16>
JwvPreferenceSettings retrieveAllPreferences.
11/7/17
workspace can read vanilla table abc; so can jas.
workspace can read j_pref; jas gets ORA-24734. 8 and 10
workspace can read abc with log raw; jas gets gpf. 8 and 10.
9/3/17
Reviewed all Jas maps. Updated using Did “Release latest apps and required maps” where necessary
and set version to 3.78a
Did “Release latest apps and required maps”
Fixed below ——————–
resetDefaultIconManagers
WkIconManager resetDefault
resetDefault
default := nil
JwgWindowBuilderExtensions ————————–
start in dir is 'C:\Users\Donald\Documents\Instantiations\VA Smalltalk\863Sep03'
make this default user my documents and put all the dlls and env stuff there.
create a desktop shortcut
wtf with gdk.cat?
9/2/17
fixed these
Error: 74 Collection»#isSequenceable cannot be added by VisualWorksCLDTCompatibility 1.0 (JAS G) because it is provided by CLDT V 8.6.3 [378].
Error: 74 SequenceableCollection»#lastIndexOf:ifAbsent: cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 SequenceableCollection»#readStream cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 SequenceableCollection»#writeStream cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 SequenceableCollection»#lastIndexOf: cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 Number»#isZero cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 Number class»#readSmalltalkSyntaxFrom: cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 Number class»#readFrom: cannot be added by VisualWorksCLDTCompatibility« 1.0 (JAS G) » because it is provided by CLDT V 8.6.3 [378].
Error: 74 Object»#isBlock cannot be added by ANLKernel 2.3.1.6, R2.0 SR0.06 because it is provided by Core V 8.6.3 [378].
Error: 74 Block»#isBlock cannot be added by VAANLKernel 2.3.0.2 because it is provided by Core V 8.6.3 [378].
Error: 74 Number»#asNumber cannot be added by ANLKernelMagnitudes 1.0, JAS C because it is provided by CLDT V 8.6.3 [378].
9/1/17
reinstall 8.6.3
copied mgr86.dat from progdata to avast\datfiles
using new environment to create 863 maps
using file i/o NOT emsrv
8/24/17
successfully packaged the jacs, but it was 4761m and the old one is 5237m.
running it with (jasR20SR321 - Chuck 7.5 !!!!!) said that the sim version was 3.3 and jacs was 3.2
(jasR20SR321 - Chuck 7.5 !!!!!) runs with the old jacs so the old jacs must be 3.3.
3.3 is set in JwvHCIJacsProxy»releaseLevelModule 3/3/2009
3.2 is set in JwvJACSServer»releaseLevelModuleBasedOn: 9/27/2006 WHY NO LATER KWAS version?
in GdkServerInterface look at validateModuleReleaseLevel: aReleaseLevel for: aClientName and validateModuleReleaseLevel: aReleaseLevel for: aClientName on: aSystemName. see also releaseLevelModule.
The mgr75.dat file I was using 635m.
Got finalmgr75.dat from cd which is 800m+ and imported all most recent maps from mgr75.dat.
Many newer versions of packaging.
updated jas dev jso to 3.21
load that
see what is missing from jas pkg sim win base and update it
start clean image
load jas dev jso
load jwxpackaging
JwvDatabaseModelsApp have SQL calls
'SELECT * FROM JWARS_Preferences WHERE OWNER IN ( DONALD , DEFAULT ) ORDER BY PREFLEVEL'
To do
find out what code we use from widgetKit, etc.
environments imagea are in users\donald\documents
data store files are in users\donald\appdata\roaming
querySpec are identical
'SELECT * FROM JWARS_Preferences WHERE OWNER IN ( DONALD , DEFAULT ) ORDER BY PREFLEVEL'
'SELECT * FROM JWARS_Preferences WHERE OWNER IN ( DONALD , DEFAULT ) ORDER BY PREFLEVEL'
AbtOracle8DatabaseConnection
3/21/17
moved lenovo to Allison VM
changed OSStructure class»alignmentOf: aTypeDescriptor at: position
PG32.DLL - Contains low level graphics routines designed to provide high performance graphing and charting facilities. The dynamic link library (DLL) is bound to the image when you start it up. This DLL is required at runtime. This file should be copied to the directory that contains your VA Smalltalk binary files.
added C:\temp\JAS\bin to the PATH dq
saveImage
"Save the image using the default name."
self saveImage: self imageFileName.
Transcript show: DateAndTime now printString;cr.
2017/03/19
jasR20SR321 - Chuck!!!!! runs a 7.5 open image in 13 minutes. initialization takes 4 minutes
JwarsRuntime»fatalError: has the word JWARS. Fix this and other occurrences.
PamMessagePrompter is not centering text properly. smartWidth is ok.
tried to load old map z.ST: Server Workbench, Base V 7.5.0 [88] JAS B. but failed with duplicate class def
Jacs has req maps jacs core and jacs nt. jacs core has req map jacs base. jacs base has req map jacs base prereqs.
Packaging using jasR20SR320 HCI (7.5) has fatal error using JAS Packaging: Jacs R2.0 SR2.00
Packaging using JAS Packaging: Jacs R2.0 DR2.23 has same fatal error as above in GdkNetworking class»#packagingRulesFor: - Does not compile because it cannot find class SstInvocationHandler which is in app SSTMethodInvocatio which is in map Server Smalltalk Kernel
Try this in Amanda VM
3/8/17
Jas9.5 in environments has 12/25 date with packaging loaded. I think it was used to create new jacs.
jasdev9.0 has latest jas dev jso loaded. will try to get the gui to come up. no jas.ini error.
There is a jas.ini in jas9.1 through jas9.4.
I left the registry pointing to the new jacs. Copying the old command line did not work. use 'sc add'
2017/02/27
-
load the Server Workbench
load your packaging rules class
Jas Packaging - Jacs has Jas Packaging - Jacs Core and Jas Packaging - Jacs Nt as required maps
Jas Packaging - Jacs Core has Distributed Kernel and Jas Packaging - Jacs Base as required maps
2017/02/15
2017/01/29
sqlplus system/jwars works which means Oracle started at Windows boot.
boot did not start a listener
hci and env live in progFiles\jas\bin
jacs.ini is in progFiles\jas\jacs
modified jacs.ini to this ServerName=JAS10
2017/01/01
db name is JAS
lives in C:\app\ora12c\oradata\JAS
sometime have to start listener even though it is set to automatic
use system/jwars to avoid as sysdba
Files ending in sql are meant to be used with sqlplus, e.g., 'sqlplus jwars/password@jas @someFile.sql'.
used expp.bat to dump the db with only tables defined.
ran chris's table creation script with log files
did imp of big exp*.dmp file and got many errors
ran chris's data import step with logs
next go back to ayeone and exp scenario and env files
created directory specified in the ayeone expdat.dmp file.
created a JAS10 db pointing to above dir.
import seemed to be ok
next rename jas10 to jas or make jas look for jas10.
2016/09/18
uninstalled 8.6.3
deleted progFile/Instantiations
got 7.5 from work
75 install fails leaving $path variables in the ini file
installer installs Jas so that it must be run as admin
commented out logging to Windows log file that was ok in XP but disallowed in Win7
put breakpoint in GdkError to find the JwvSimRunHistorySOmething was missing
todo: dump all tables belonging to JWARS in Ayeone; then compare to Lenovo.
Scripts are in Oracle/new on Lenovo. DONE
put breakpoint in illuminationGridAt: and see why SEA(15s-15N, 90E-150e) is missing. No areas at all.
-
-
From 64 bit oracel 12c install: Your database configuration files have been installed in C:\app\ora12c while other components selected for installation have been installed in C:\app\ora12c\product\12.1.0\dbhome_1. Be cautious not to accidentally delete these configuration files.
ora12c/ora12c is the oracle windows user
the ayeone oracle is 10.2.0.1.0
2013 08 31
jas mysql New has no Oracle
JAS Dev: JSO R2.0 SR3.20 in 320a JAS ss2011 Odbc V2_2 Med image
installed Active Perl
-
ran cpan from the command lie; it downloads mingw and make
-
look at readme in \ora2pg
env_province_grid is 1.2M
tl_location_to_track_grid 2M
tl_track_to_loss_curve 5M
tranmission_loss 1.9M
2015 05 15
Env tblspace use is 7.9 of 10 gig
index tblspace is 100m of 1 gig
instruments tblspace is 2.5 of 8 gig
jar001 1 m of 2 gig
primary is 100m of 2 gig
scenarioComponents is 666m of 2 gig
There are two advapi32.dll, the 64bit version is in C:\Windows\System32, the 32bit version is in C:\Windows\SysWOW64. you may check whether they exist in your system.
To find out which dll brings the error, you can use fuslogvw.exe and DependencyWalker.exe to find out which dll is missed.
-
HKEY_LOCAL_MACHINE
SYSTEM
CurrentControlSet
Services
Eventlog
Application
Security <----------------------------- write is restricted in Win7
System