Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by John Hurle » Sat, 12 Mar 2011 12:46:05


Bill:

I would start looking here ...

ERROR org.openbravo.data.Sqlc - SQL error in query: SELECT
B.C_DEBT_PAYMENT_ID, L.NAME AS PAYMENTRULE, BP.NAME AS BPARTNER,

and here ...

[java] java.sql.SQLRecoverableException: No more data to read from
socket

Sounds like a bad query somehow ... can you try running that query
thru some other tool ( like SQL Developer or Toad etc ) to give you
some idea on what needs to be check out and/or fixed?
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Tue, 15 Mar 2011 08:09:25


To give myself another environment I am putting together a 10g Rel 2
environment on a Red Hat EL system.

My initial test will be to see if it will build a database from the
application build on VMS via JDBC and directing the host to this new
environment. If it works then I am a bit ahead of where I was and I
can "encourage" Oracle to look at the issue.

If it still fails then I have one more option of doing application
build on the RHEL environment and see where that gets me.

Honestly, I am skeptical at the possibility that the query is "bad" or
"malformed" and more suspicious of the ACCVIO on the NETV2 server
process.

What ever way I end up if I can get the database built and the
application built, on VMS preferably, then I am ahead and can then
start looking at modifying to use the eventual target. Likewise if I
can get it built on some other environment then potentially I can port
the DB itself to OpenVMS as a start.

All part of learning...

Thanks!

Bill.

 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by John Hurle » Wed, 16 Mar 2011 02:19:03

Bill:

# Honestly, I am skeptical at the possibility that the query is "bad"
or "malformed" and more suspicious of the ACCVIO on the NETV2 server
process.

Being skeptical is always a good thing.

It sure looks like you have an error message in your output. That
error message appears to give a specific SQL query as causing a
problem ... with "no more data" possibly involved.

While you are doing other setup and other testing can you isolate the
specific query and prove it good ( or bad )?
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by joel garr » Wed, 16 Mar 2011 03:11:27


Last time I was on VMS (which is like 16 years ago), it was notable
for doing things like giving you an access violation for not having
enough privilege on something obscure, so the support joke was to tell
people to increase their quota, which would magically solve many
problems, including not following installation directions, plus they'd
have to get the privs needed to increase quota to get a big enough
working set.

So for Bill's situation, it could be something like, the java asks for
too much of something, the OS says "screw you," the openbravo says
"oh, the sql didn't work," reporting the accvio and a sql didn't work
error - which was that it didn't get any data.

Wild speculation, of course, an important thing is to check the sql
elsewhere as you suggested, though it can be a real pain to get the
context right with this kind of software stack. VMS has plenty of
instrumentation, but that needs plenty of specialized training.
>> >> While you are doing other setup and other testing can you isolate the >> specific query and prove it good ( or bad )?

jg
@home.com is bogus.
"my money has already been printed"
http://www.yqcomputer.com/
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Thu, 24 Mar 2011 05:41:41


Ok, finally got my Oracle 10gR2 Linux (x86) implementation up and
running and it gets past the offending point in the code.

Configuration is:

ANT and necessary code residing on OpenVMS I64 V8.4 with Oracle JDBC
Thin Driver pointing at the Linux DB instance.

Linux environment is Red Hat V4.8 with Oracle 10gR2.

In this environment the ANT build of OpenBravo ERP is proceeding... I
has proceeded past the point of failure when entirely on OpenVMS.

So this points at the NETV2 process as either not having sufficient
quotas or similar. I would generally doubt privileges as well over
15000 SQL transactions had already been processed and additionally
other transactions of a similar nature have been executed. I am
leaning toward the quotas as the particular transaction is fairly
large but not to the extent of surpassing SQL limits as it has run
properly in this other environment.

Any suggestions?

Any place I should look for the particular quotas the NETV2 server is
started with and possibly how to change?

Now, that said, I have now seen an error - farther on, where is it
JAVA and fork oriented... So I will need to explore that problem now
while I think of the other issues and talk with folks.

Thanks,

Bill.
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by joel garr » Thu, 24 Mar 2011 08:48:30


I take it you are likely more familiar with MOS OpenVMS: How Process
Quotas are Set for the SQL*NET listener [ID 68226.1] than I am. But I
point there because it shows how to get quota anomalies.

jg
@home.com is bogus.
http://www.yqcomputer.com/
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Thu, 24 Mar 2011 22:51:52


Thanks, Joel. I think I have found what I need in the Admin/OpenVMS
manual. There is a discussion of quotas and logical names in that
book. The only issue is whether they are actually picked up the way
it is documented or if there are other issues I need to look at to
make it work.

In the mean time I have a Java memory issue not related to this I need
to resolve to get all the way through the build using the Linux RDBMS
server first. Then I will concern myself with this issue.

But first I need to make some money...

Regards,

Bill.
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by Syltre » Sat, 26 Mar 2011 02:51:48

"BillPedersen" < XXXX@XXXXX.COM > wrote in message
news: XXXX@XXXXX.COM ...
On Mar 10, 11:46 pm, John Hurley < XXXX@XXXXX.COM > wrote:

I have these quotas setup on my Oracle server, and they are based upon
Oracle's recommendations (among others, I run lots of different things on
that server).

SYSGEN> SHOW PQL_M
Parameter Name Current Default Min. Max. Unit
Dynamic
-------------- ------- ------- ------- ------- ---- -------
PQL_MASTLM 2048 4 4 -1 Ast
D
PQL_MBIOLM 1000 4 4 -1 I/O
D
PQL_MBYTLM 1000000 128000 128000 -1 Bytes
D
PQL_MCPULM 0 0 0 -1 10Ms
D
PQL_MDIOLM 2048 4 4 -1 I/O
D
PQL_MFILLM 2048 2 2 -1 Files
D
PQL_MPGFLQUOTA 2000000 512000 512000 -1 Pagelets
D
internal value 125000 32000 32000 -1 Pages
D
PQL_MPRCLM 20 0 0 -1 Processes
D
PQL_MTQELM 2000 0 0 -1 Timers
D
PQL_MWSDEFAULT 43888 16384 16384 -1 Pagelets
internal value 2743 1024 1024 -1 Pages
PQL_MWSQUOTA 87776 32768 32768 -1 Pagelets
D
internal value 5486 2048 2048 -1 Pages
D
PQL_MWSEXTENT 1873920 65536 65536 -1 Pagelets
D
internal value 117120 4096 4096 -1 Pages
D
PQL_MENQLM 8192 64 64 -1 Locks
D
PQL_MJTQUOTA 8192 0 0 -1 Bytes
D

HTH
Syltrem


 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Sat, 26 Mar 2011 23:09:21


....

Thanks, will see if they get me anyplace. Unfortunately in looking at
the Oracle manuals and the command procedures and the comparing them
to the actual running ORASRV_NETV2 process' quotas they do not seem to
give me a mechanism that I have been able to isolate to affect the
PGFLQUOTA for the server. It is still stuck at 125000. I seem to
have been able to affect other quota items for the process with the
changes to the PQL_M* values you provided.

Will run some more tests and see if I can glean where this quota is
controlled since I suspect it is really the one I need to change.

Regards,

Bill.
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Sun, 27 Mar 2011 00:20:42


On further examination I think I have found the relationship.
Building the test case now... Will get back later..

Bill.
 
 
 

Oracle 10g Release 2 OpenVMS I64 V8.4 porting OpenBravo ERP

Post by BillPeders » Sun, 27 Mar 2011 02:22:45

ell, someone with access to source is probably going to need to look
at this...

I tripled the size of PGFLQUOTA and still not luck. I also adjusted
all the other parameters as per your suggestions (with Oracle's help
to you...).

I still get an ACCVIO at the exact same location:

%TRACE-I-END, end of TRACE stack dump
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
address=000000011036DA88, PC=000000000A1F66A0, PS=0000001B
%TRACE-F-TRACEBACK, symbolic stack dump follows
image module routine line rel PC
abs PC
ORACLE lnxnur lnxnur 90029 0000000000000130
000000000A1F66A0
ORACLE kke kkenrm_internal 1507253 000000000002AE22
00000000044CB4F2
ORACLE kke kkenrms 1507446 000000000002C472
00000000044CCB42
ORACLE kke kkegvl 1507822 000000000002DF22
00000000044CE5F2
ORACLE kke kkepsl 1505613 000000000001A6A2
00000000044BAD72
ORACLE kke kkecdn 1504319 0000000000008C12
00000000044A92E2
ORACLE kko kkotap 1593625 00000000000E3732
000000000479EEE2
ORACLE kko kkoiqb 1587290 000000000007E8F2
000000000473A0A2
ORACLE kko kkooqb 1584126 0000000000047FA2
0000000004703752
ORACLE kko kkoqbc 1582340 000000000001EA12
00000000046DA1C2
ORACLE apa apakkoqb 1429628 0000000000007E22
00000000021BA002
ORACLE apa apaqbd 1429900 0000000000009822
00000000021BBA02
ORACLE kkqct kkqctCostTransfQB 1456834 0000000000007EB2
0000000004A1F472
ORACLE kkqct kkqctdrvSU 1460765 0000000000025322
0000000004A3C8E2
ORACLE nsoqbc nsotruns 1490081 0000000000002AC2
0000000006F62772
ORACLE nsoqbc nsoqbc 1489958 0000000000001CE2
0000000006F61992
ORACLE kkqct kkqctdrvTD 1462400 0000000000042682
0000000004A59C42
ORACLE kkqs kkqdrv 1517685 0000000000001BA2
0000000004A617F2
ORACLE kkqct kkqctdrvIT 1462309 0000000000041682
0000000004A58C42
ORACLE apa apadrv 1428430 0000000000002162
00000000021B4342
ORACLE opitca opitca 1545403 00000000000041D2
0000000006FDDBC2
ORACLE kks1 kksFullTypeCheck 1511740 0000000000000422
0000000004FB3162
ORACLE rpi rpiswu2 1429092 0000000000004812
0000000001B52182
ORACLE kks1 kksSetBindType 1515485 0000000000026F42
0000000004FD9C82
ORACLE kks kksfbc 1531795 0000000000019B52
0000000004FA2B22
ORACLE opiexe opiexe 1575172 0000000000005632
00000000010A0F12
ORACLE kpoal8 kpoal8 1463220 0000000000002ED2
0000000005982F82
ORACLE opiodr opiodr 1504090 0000000000002212
00000000010E16D2
ORACLE ttcpip ttcpip 412801 0000000000005562
000000000B17E252
ORACLE opitsk opitsk 1527361 0000000000002512
000000000110B5A2
ORACLE opiino opiino 1453303 0000000000004CD2
0000000006FB2852
ORACLE opiodr opiodr 1504090 0000000000002212
00000000010E16D2
ORACLE opidrv opidrv 1450763 0000000000002F62
00000000010DBCF2
ORACLE sou2o sou2o 1405727 0000000000000202
0000000001BA2662
ORACLE opimai main 1402531 0000000000000942
0000000000080BE2
ORACLE opimai __main