Could not prepare resource 'OatmialResource

Question

We´ve been facing a problem while weblogic asks tuxedo to prepare XA transaction before commit, as no response is received it calls rollback after all. Would anybody provide feedback on this error ?
Environment details: Weblogic 10.3.2 acessing Tuxedo 8.1 through WTC
Exception occurred during commit of transaction Xid=BEA1-0441E759B8D59E472B1E(3521575),Status=Rolled back. [Reason=javax.transaction.xa.XAException],numRepliesOwedMe=0,numRepliesOwedOthers=0,seconds since begin=30,seconds left=30,XAServerResourceInfo[WLStore_DefaultDomain__WLS_DefaultServer]=(ServerResourceInfo[WLStore_DefaultDomain__WLS_DefaultServer]=(state=rolledback,assigned=DefaultServer),xar=WLStore_DefaultDomain__WLS_DefaultServer18552479,re-Registered = false),XAServerResourceInfo[jdbc_xxx_DefaultDomain]=(ServerResourceInfo[jdbc_xxx_DefaultDomain]=(state=rolledback,assigned=DefaultServer),xar=jdbc_xxx,re-Registered = false),XAServerResourceInfo[OatmialResource]=(ServerResourceInfo[OatmialResource]=(state=rolledback,assigned=DefaultServer),xar=com.bea.core.jatmi.internal.TuxedoXA@1d3b2e9,re-Registered = false),SCInfo[DefaultDomain+DefaultServer]=(state=rolledback),local properties=({weblogic.jdbc.jta.jdbc_xxx=[ No XAConnection is attached to this TxInfo ]}),OwnerTransactionManager=ServerTM[ServerCoordinatorDescriptor=(CoordinatorURL=DefaultServer+X.X.X.X:XXXX+DefaultDomain+t3+, XAResources={OatmialResource, WLStore_DefaultDomain__WLS_DefaultServer, jdbc_xxx_DefaultDomain, WSATGatewayRM_DefaultServer_DefaultDomain},NonXAResources={})],CoordinatorURL=DefaultServer+X.X.X.X:XXXX+DefaultDomain+t3+): weblogic.transaction.RollbackException: Could not prepare resource 'OatmialResource
at weblogic.transaction.internal.TransactionImpl.throwRollbackException(TransactionImpl.java:1848)
at weblogic.transaction.internal.ServerTransactionImpl.internalCommit(ServerTransactionImpl.java:339)
at weblogic.transaction.internal.ServerTransactionImpl.commit(ServerTransactionImpl.java:233)
Caused by: javax.transaction.xa.XAException
at com.bea.core.jatmi.internal.TuxedoXA.internalPrepare(TuxedoXA.java:625)
at com.bea.core.jatmi.internal.TuxedoXA.prepare(TuxedoXA.java:521)
at weblogic.transaction.internal.XAServerResourceInfo.prepare(XAServerResourceInfo.java:1295)
at weblogic.transaction.internal.XAServerResourceInfo.prepare(XAServerResourceInfo.java:500)
at weblogic.transaction.internal.ServerSCInfo$1.run(ServerSCInfo.java:361)
... 3 more

Answer 1

Thanks for looking at this. Here is the tuxedo configuration.
#### DMCONFIG
wtc_xxxxxx
GWGRP=WTC
TYPE=TDOMAIN
DOMAINID="wtc_xxxxxx"
BLOCKTIME=20
DMTLOGDEV="/opt/xxxxx/log/xxxxxUDL"
DMTLOGNAME="WTC_DMTLOG"
DMTLOGSIZE=1024
BLOB_SHM_SIZE=5000000
CONNECTION_POLICY=ON_DEMAND
wtc_xxxxxx NWADDR="//x.x.x.x:xxxx"
E0000XXXX
AUTOTRAN=N
CONV=N
API=ATMI
LDOM=xxxxxx
RDOM=mainframe
RNAME="X000:XXXXXXX"
INBUFTYPE="STRING"
OUTBUFTYPE="STRING"
TRANTIME=120
FUNCTION=DPL
#### UBBCONFIG
TLOGDEVICE="/opt/xxxxx/log/xxxxxUDL"
TLOGNAME="TLOG"
TLOGSIZE=1024
xxxxxx LMID=SITE1 GRPNO=777
TMSNAME=oratmss TMSCOUNT=10
OPENINFO="Oracle_XA:Oracle_XA+Acc=P/exxxxxx/xxxxxx+SesTm=240+SesWt=60+DbgFl=0x7+SqlNet=xxxxxx+LogDir=/opt/xxxxx/log"
GWADM SRVGRP=WTC SRVID=666
GWTDOMAIN SRVGRP=WTC SRVID=888
xxxxxxxxxxxx SRVGRP=xxxxxx SRVID=999
REPLYQ=Y RQADDR=RQxxxxxxxxxxxx
RESTART=Y MAXGEN=10 GRACE=0 MIN=1 MAX=10
CLOPT="-A -- -x"
Here is the tuxedo error. It happens before my call from weblogic to tuxedo, not a 100 % sure it´s related:
XXX.081011:120404.XXX!GWTDOMAIN.15430.1.0: gtrid x0 x4e42592e x35f66: LIBTUX_CAT:4055: ERROR: No service '..GWA9' in group 'WTC'
XXX.081011:120404.XXX!GWTDOMAIN.15430.1.0: gtrid x0 x4e42592e x35f66: LIBGW_CAT:5241: WARN: Can't send log request to GWADM. Performance may suffer
Inside CRM log:
When everything is ok:
REQUEST FOR Verb=SNACRM_PREPARE
==> Sending socket data: <<==
When there is an error:
REQUEST FOR Verb=SNACRM_PREPARE
==>> Socket data received: <<==
Couple of additional info:
- The error is intermittent, but when sna is "restarted" everything works just fine up to the error happens again.
- Mainframe is processing just fine
- Oracle is processing just fine
- The error happens when weblogic calls tuxedo via WTC within a XA transaction. Weblogic asks oracle (the other xa participant) to prepare and receives ok, when it asks tuxedo and no response is received 30 secods later (my weblogic transacion timeout) it calls rollback. When I increase this transaction time the same situation happens but the call to rollback happens x seconds later. Rollback is processing just fine.
- When WTC debug was enabled nothing different from this information has been found.
- Based on the research I´ve been doing it seems it would be related to the items below, but as we are not sure, still looking at other options while the ticket is processed:
http://download.oracle.com/docs/cd/E13203_01/tuxedo/eam/v91/relnot/relnot.html
CR233501 & CR235337
Fix for outbound transactions getting lost in CRM after SNACRM_PREPARE is sent to CRM.
https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&doctype=HOWTO&id=776644.1
Thanks much.