Error using repotools-jazz -export
![]()
Dear community,
We have just upgraded our solution from version 2.x to 3.0.1.2. At the moment I am working on our nightly backup procedure, and I am following the steps indicated on the "boulder" help web site. We have a distributed topology using 3 servers; one for CCM, one for QM and one for JTS. When it comes to running "repotools-jazz -export" for the QM installation, we get an error when validating the export of "TestSuiteExecutionRecord": 2012-01-30 17:09:42,894 Validating exported items for "com.ibm.rqm.execution"... .... 2012-01-30 17:09:42,896 Finished validating exported items in 3ms. This is pretty much the only error indication in the log file. What can you tell me about it, to work towards a solution? Best, Martin |
7 answers
![]()
Ralph Schoon (63.3k●3●36●46)
| answered Feb 06 '12, 8:30 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER
Hi Martin,
as far as I know repotools -export is only meant to be used if there is a need to migrate from one DB vendor to another or during migration if required in the steps. The repotools import/export performance does not scale for big deployments. You should use file backup and Database backup techniques to back up. Please see https://jazz.net/library/article/547 for how to backup 3.0.x based CLM environments. The article only covers RTC so far, but if you replicate it three times (assuming you use RRC) and remove the backup of the RRC DB which is in JTS you should be fine. Make sure to backup the index files wile stopping the indexer. If the indexes are missing entries after restore they will be recreated. Look for the backup order between JTS and the other DB's or run an offline backup for all data over night. Since your approach also requires the DB's to be offline you should probably go for an offline backup first and take all servers offline in parallel. Which DB are you using? With respect to your errors is there anything in the logs? You might want to contact support to get access to the online verification tools. |
![]() as far as I know repotools -export is only meant to be used if there is a need to migrate from one DB vendor to another or during migration if required in the steps. The repotools import/export performance does not scale for big deployments. The reason we are using the repotools, is that it is the way indicated here http://publib.boulder.ibm.com/infocenter/clmhelp/v3r0m1/index.jsp?topic=%2Fcom.ibm.jazz.repository.web.admin.doc%2Ftopics%2Ft_jfs_backup.html. So for each of the servers, we are backing up the Jazz and Tomcat config files indicated on that page, and doing repotools -export. The article only covers RTC so far, but if you replicate it three times (assuming you use RRC) and remove the backup of the RRC DB which is in JTS you should be fine. We are not using RRC. We have CCM, QM and JTS on 3 machines, so we would be backing those up. Make sure to backup the index files wile stopping the indexer. If the indexes are missing entries after restore they will be recreated. We have never gotten the "repotools -suspendIndexer" to work, even though it is indicated as a step on the page I linked above. The error we get is: CRJAZ1357E Failed to login into the server 'https://localhost:9443/jazz'. Check the login credentials. Look for the backup order between JTS and the other DB's or run an offline backup for all data over night. Since your approach also requires the DB's to be offline you should probably go for an offline backup first and take all servers offline in parallel. Is it important that all servers are offline when we do the backup? At the moment, all servers are scheduled to perform backup individually at 02:00 am, using a script started with a cron job. We are using DB2-ExpressC database. Should we just be backing up the configuration files, and doing a DB2 backup of the databases instead? With respect to your errors is there anything in the logs? You might want to contact support to get access to the online verification tools. The logs say nothing except what I pasted to you already. Thanks for your time. - Martin |
![]()
Ralph Schoon (63.3k●3●36●46)
| answered Feb 06 '12, 11:05 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER as far as I know repotools -export is only meant to be used if there is a need to migrate from one DB vendor to another or during migration if required in the steps. The repotools import/export performance does not scale for big deployments. The reason we are using the repotools, is that it is the way indicated here http://publib.boulder.ibm.com/infocenter/clmhelp/v3r0m1/index.jsp?topic=%2Fcom.ibm.jazz.repository.web.admin.doc%2Ftopics%2Ft_jfs_backup.html. So for each of the servers, we are backing up the Jazz and Tomcat config files indicated on that page, and doing repotools -export. You can do that, but I would advise not to. If you do, try a restore and look at the cycle time. If it starts exceeding half a day you might want to look into backing up the database instead. The article only covers RTC so far, but if you replicate it three times (assuming you use RRC) and remove the backup of the RRC DB which is in JTS you should be fine. We are not using RRC. We have CCM, QM and JTS on 3 machines, so we would be backing those up. Make sure to backup the index files wile stopping the indexer. If the indexes are missing entries after restore they will be recreated. We have never gotten the "repotools -suspendIndexer" to work, even though it is indicated as a step on the page I linked above. The error we get is: CRJAZ1357E Failed to login into the server 'https://localhost:9443/jazz'. Check the login credentials. The error message clearly points to missing login credentials. If in doubt I run a repotools without parameters, pipe the output to a text file and save it somewhere and examine the commands. Examine suspendIndexer and look for how to provide a user and password with JazzAdmin repository permissions. You also might want to provide the repository URL to contact to, since I assume https://localhost:9443/jazz is not the URL you are using. Look for the backup order between JTS and the other DB's or run an offline backup for all data over night. Since your approach also requires the DB's to be offline you should probably go for an offline backup first and take all servers offline in parallel. Is it important that all servers are offline when we do the backup? At the moment, all servers are scheduled to perform backup individually at 02:00 am, using a script started with a cron job. We are using DB2-ExpressC database. Should we just be backing up the configuration files, and doing a DB2 backup of the databases instead? Please carefully read the article I mention. There s an importance in the sequence to some extend. JTS writes data into the other applications. If all servers are down, no problem. Maybe take JTS down first. Please read the article for more. I would suggest to use a DB2 backup rather than repotools. With respect to your errors is there anything in the logs? You might want to contact support to get access to the online verification tools. The logs say nothing except what I pasted to you already. Thanks for your time. - Martin I'd get in contact with support. Offline and online verification tools might point to what the issue is. |
![]()
Ralph Schoon (63.3k●3●36●46)
| answered Feb 07 '12, 7:09 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER
Hi,
I checked http://publib.boulder.ibm.com/infocenter/clmhelp/v3r0m1/topic/com.ibm.jazz.repository.web.admin.doc/topics/t_jfs_backup.html and it mentions it is possible to use the repotools command for backup. However, it mentions also DB specific tools. Experience shows the repotools, especially the import is slow compared to a database backup. So I would rather use a database backup. |
![]() Hi, Hi Ralph, Thanks for your response! I must admit, I believe, this was a case of when you read about one way to do something you tend to forget there were ever another way :) |
![]()
Hi,
Having read through this article, there is just one question that pops to mind. I have looked through the teamserver.application files, and only one of the values mentioned for index files (com.ibm.team.fulltext.indexLocation) is present, and the value here is already backed up, as it is a sub directory of /server/conf/<app>. As we do offline backup, I see there is no need to suspend indexing either. So when we restore, we simply restore the indexing files as part of the /server/conf/<app> configuration restore too. Generally, I think its a very good article, and I think perhaps it should be linked from the "boulder" site! |
![]()
Ralph Schoon (63.3k●3●36●46)
| answered Feb 10 '12, 4:59 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER Hi, Hi Ralph, Thanks for your response! I must admit, I believe, this was a case of when you read about one way to do something you tend to forget there were ever another way :) Martin, we are all state-full beings 8-). I created https://jazz.net/jazz/web/projects/Rational%20Team%20Concert#action=com.ibm.team.workitem.viewWorkItem&id=195625 to improve the help topic and make the data base backup more and repotools less prominent. |
Your answer
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.