How long do repotools -rebuildIndices take to run?
John Owen (46●24●22)
| asked Jan 15 '13, 4:04 a.m.
retagged Jul 08 '13, 11:40 a.m. by Jennifer Cianchetta-Riordan (251●2)
We've just upgraded a repository to Jazz V4.0.0.1 and the diagnostics are displaying:
|
Accepted answer
Chances are your 4001 upgrade install location just can't find the index locations of the previous installation. We tried to provide information in the interactive upgrade guide that allowed you to be more cognizant of your existing index location prior to upgrade so such that during the upgrade, the indices would either 1) remain in their existing 'stable' locations and be accurately pointed at from the teamserver.properties or 2) be copied from their existing locations to a new location specified in the 'new' teamserver.properties file of the upgraded software.
Just look for this line in both your JTS and <app-specific> teamserver.properties, where app is ccm or qm: com.ibm.team.fulltext.indexLocation. Here is an excerpt from the v4.0.1 infocenter on verifying your index locations prior to upgrading: Before you start the upgrade, verify that the index locations in the application's properties files are pointing to absolute paths on the file system, rather than relative paths.
John Owen selected this answer as the correct answer
Comments Thank you for your append Rosa.
1
The v4.0.1 infocenter has a topic for each of these commands.
|
4 other answers
I just did it a week ago. On JTS:
On CCM:
On QM:
Sizing info of my environemnt:
Remark:
Comments Thanks for sharing Guido.
1
Guido Schneider
commented Jan 15 '13, 5:47 a.m.
your welcome. Hope to earn some reputation points. This year goal is to be over 2000 and on first page of the users list. :-) |
Hello
I'm involved in a RTC 4.0.6 migration project, these are the times of the repotools commands: Hardware specifications: WAS 8.5.5 + DB2 10.1 Server RHEL 6.5 / 64 bits 8GB RAM 2 CPU (4 cores) Pentium XEON 50110 1.6GHz DD.BB Sizing: JTS DB about 11GB CCM DB about 48GB On JTS: ./repotools-jts.sh -reindex --> 1 min. ./repotools-jts.sh -rebuildTextIndices --> 30 sec. ./repotools-jts.sh -rebuildIndices --> 30 sec. On CCM: ./repotools-ccm.sh -reindex --> 55 min ./repotools-ccm.sh -rebuildTextIndices --> 40 min. ./repotools-ccm.sh -rebuildIndices --> 15 min. |
Ralph Schoon (63.5k●3●36●46)
| answered Jan 15 '13, 4:27 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER edited Jan 15 '13, 5:35 a.m.
Hi John,
please be careful which index command you run. One is creating the index on the database tables and typically not intended to be run by a user (I have been told). We are in the process of updating articles (the backup articles) and creating new content, based on recent information we gathered about the indexing story. I can't give you an idea about how long it takes to recreate the indexes. I think you might have to try. It would be good if you measure how long it takes. Here a summary of reindex commands with our latest information: Recreate the Fulltext IndexYou can recreate the Fulltext index files using the command: repotools-${APP} -rebuildTextIndices Recreate the JFS IndexYou can recreate the JFS index files using the command: repotools-${APP} -reindex all Comments
Ralph Schoon
commented Jan 15 '13, 4:29 a.m.
| edited Jan 15 '13, 4:30 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER
I just realized you are missing the database indexes. I am unsure how that could have happened. I would suggest to talk to support to find out how that could have happened and be prevented to happen in the future.
John Owen
commented Jan 15 '13, 7:08 a.m.
Thanks for your reply, Ralph. All we need to do is rebuild the the indices using the -rebuildIndices option of the repotools command. I've done some further exploring. Sam had to do this on another repository last weekend it took about 15 minutes for the ccm application which comprised a 35GB database.
1
John Owen
commented Jan 15 '13, 12:26 p.m.
The -rebuildIndices has completed on the 70GB /ccm database and it took about 30 minutes to run.
Hi John, thanks for the update. I am still curious how comes the database indexes got messed up. Did anyone work directly on the database and changed policies or maintenance settings? Was this caused by an issue with the database? I just want to make sure this is not a symptom of a lingering problem somewhere.
|
My two cents - our RTC DB has about 35 GB data and rebuildIndices took about 30 mins. Because it is just creating DB indexes, it won't take as much time as rebuildTextIndices which usually took hours.
|
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.