| Author |
Message
|
| j.f.sorge |
 Posted: Fri Sep 30, 2011 6:26 am Post subject: cross platform migration (Solaris to Linux) |
 |
|
Master
Joined: 27 Feb 2008
Posts: 218
|
Currently we are running WMB 6.1 on Sun Solaris and want to switch to Linux platform without redeploying all our flows. So we tried to do the following steps
create backup of broker database, /var/mqsi/components, config manager, queue manager definitions
create clone of broker database
create new queue manager with old name and definitions from backup
create new config manager with old name and restore backup
create new broker with old name into clone of database and restore /var/mqsi/components directory
When using mqsilist on Linux box it shows the same configuration the Sun Solaris box had.
So we tried to start the ConfigManager on Linux box and it completed successfully.
After that we tried to start the Broker but each execution group fails to start with the following error messages
| Code: |
Sep 30 10:28:43 hostname WebSphere Broker v6109[25040]: (broker)[1]BIP8873I: Starting the component verification for component 'broker'. : broker.mqsistart: /build/S610_P/src/CommandLine/ImbCmdLib/ImbComponentVerification.cpp: 74: ImbComponentVerification::verify: :
Sep 30 10:28:45 hostname WebSphere Broker v6109[25049]: (broker)[1]BIP2001I: The WebSphere Message Brokers service has started at version 6109; process ID 25054. : broker.service: /build/S610_P/src/AdminAgent/ControlProcess/Linux/ImbControlService.cpp: 345: ImbControlService::StartNewAA: :
Sep 30 10:28:46 hostname WebSphere Broker v6109[25087]: (broker.executionGroupLabel)[1]BIP2208I: Execution group (64) started: process '25087'; thread '1'; additional information: brokerName 'broker' (operation mode 'enterprise'); executionGroupUUID 'executionGroupUUID'; executionGroupLabel 'executionGroupLabel'; defaultExecutionGroup 'true'; queueManagerName 'qMgr'; trusted 'false'; dataSourceName 'dataSourceName'; userId 'userId'; migrationNeeded 'false'; brokerUUID 'brokerUUID'; filePath '/opt/ibm/mqsi/6.1'; workPath '/var/mqsi'; ICU Converter Path ''. : broker.executionGroupUUID: /build/S610_P/src/DataFlowEngine/ImbMain.cpp: 539: ImbMain::start: :
Sep 30 10:28:48 hostname IBM Java[25087]: JVMDUMP032I JVM requested System dump using '/tmp/core.20110930.102848.25087.0001.dmp' in response to an event
Sep 30 10:28:48 hostname kernel: DataFlowEngine[25127]: segfault at 0000000000000000 rip 00002aaaafab6c2b rsp 00007fffd5ff4880 error 6
Sep 30 10:28:48 hostname IBM Java[25087]: JVMDUMP032I JVM requested Snap dump using '/tmp/Snap.20110930.102848.25087.0002.trc' in response to anevent
Sep 30 10:28:48 hostname IBM Java[25087]: JVMDUMP032I JVM requested Java dump using '/tmp/javacore.20110930.102848.25087.0003.txt' in response to an event
Sep 30 10:28:49 hostname WebSphere Broker v6109[25054]: (broker)[4]BIP2060W: The broker has detected that the Execution Group executionGroupLabel, process ID 25087, has shutdown. : broker.agent: /build/S610_P/src/AdminAgent/ImbAdminAgent.cpp: 4583: ImbAdminAgent::startAndMonitorADataFlowEngine: :
|
Did anyone try a cross platform migration, too?
[edit]
A short search brought the following link which mentions that a redeploy would be better when trying to switch platforms, but probably anyone has experience using the backup-way?!
[/edit]
_________________
IBM Certified Solution Designer - WebSphere MQ V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
 |
| smdavies99 |
| Posted: Fri Sep 30, 2011 7:00 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
While the Hursley guys are banging their collective heads against a wall, I have to say that what you are doing is....
Not supported for V6.1
I'm not sure about 7.0
AND
This is not how I'd do it.
Why don't you want to redploy your flows? Could it be perchance that some of the .bar files have been mislaid?
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| marko.pitkanen |
| Posted: Fri Sep 30, 2011 7:11 am Post subject: |
|
|
Chevalier
Joined: 23 Jul 2008
Posts: 440
Location: Jamsa, Finland
|
Hi,
Did you configure all data sources on Linux box likewise it was on Solaris?
--
Marko |
|
| Back to top |
|
|
| j.f.sorge |
| Posted: Mon Oct 03, 2011 11:15 pm Post subject: |
|
|
Master
Joined: 27 Feb 2008
Posts: 218
|
| smdavies99 wrote: |
| Why don't you want to redploy your flows? Could it be perchance that some of the .bar files have been mislaid? |
We tried to do the migration in order to save deployment time, but as deployment seems to be really fast on the new machine, re-deployment would be the best solution. Bar files will be restored from CVS server.
| marko.pitkanen wrote: |
| Did you configure all data sources on Linux box likewise it was on Solaris? |
Not sure. I'll have to ask the admin.
_________________
IBM Certified Solution Designer - WebSphere MQ V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
|
| j.f.sorge |
| Posted: Wed Oct 26, 2011 10:31 pm Post subject: |
|
|
Master
Joined: 27 Feb 2008
Posts: 218
|
| j.f.sorge wrote: |
| marko.pitkanen wrote: |
| Did you configure all data sources on Linux box likewise it was on Solaris? |
Not sure. I'll have to ask the admin. |
All relevant data sources had been configured correctly.
_________________
IBM Certified Solution Designer - WebSphere MQ V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
|
| marko.pitkanen |
| Posted: Tue Nov 08, 2011 2:35 am Post subject: |
|
|
Chevalier
Joined: 23 Jul 2008
Posts: 440
Location: Jamsa, Finland
|
Hi,
We had just similar segmentation fault problem and the cause was that one flow consumed too much stack. Have you checked your java heap and stack sizes ?
--
Marko |
|
| Back to top |
|
|
| j.f.sorge |
| Posted: Thu Nov 10, 2011 12:27 am Post subject: |
|
|
Master
Joined: 27 Feb 2008
Posts: 218
|
| marko.pitkanen wrote: |
Hi,
We had just similar segmentation fault problem and the cause was that one flow consumed too much stack. Have you checked your java heap and stack sizes ?
--
Marko |
Thanks for this hint! I did neither check nor change any heap / stack value on either solaris or linux. But as redeployment of all flows let them run correctly I don't think the error occured because of heap / stack configuration here.
_________________
IBM Certified Solution Designer - WebSphere MQ V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.0
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
|
|
|
