[Building Sakai] issues with contemt-runconversion.sh in 2.6.x

[Robert Cohen]

I've been trying the use the content conversion script in
content/content-runconversion.sh in a 2.6.x instance that was upgraded from
2.4.x.

Initially I had some issues with it being unable to load class
Type1BlobResourcesConversionHandler.

I tracked that down to a issue where the script was trying to load
sakai-kernel-impl-$KERNELVERSION.jar from the maven repository.

For some reason in 2.6.x it appears that this jar doesn't end up in the
maven repository. But it does get deployed successfully to tomcat. How it
gets there without a copy being squirrelled away in maven I have no idea.
Truly the ways of maven are mysterious :-).

Anyway, I modified the script to load the jars out of tomcat instead of from
the maven repository. Which seems a far more sensible thing to do to me
anyway... Of course, it does mean the script is more sensitive to updates to
the version numbers of the jars involved.
I attach the updated version in case its useful to someone else.

If this is considered generally useful and survives review, it might be
worth replacing the current 2.6.x version... And while I'm on the subject,
the top level readme.txt is getting more and more out of date..


Anway now the script now runs successfully.
But I've run it though our dev server. But I started getting errors of this
form

2010-04-29 15:51:29,894  WARN Timer-1
org.sakaiproject.search.optimize.impl.OptimizeIndexOperation - Failed to
compete Optimize 
org.sakaiproject.search.journal.api.JournalErrorException: Failed to open
permanent Index 
at 
org.sakaiproject.search.journal.impl.JournaledFSIndexStorage.getPermanentInd
exWriter(JournaledFSIndexStorage.java:1320)
.
.
Caused by: org.apache.lucene.store.LockObtainFailedException: Lock obtain
timed out: SimpleFSLock@/usr/local/tomcat/sakai/indexwork/index/write.lock
        at org.apache.lucene.store.Lock.obtain(Lock.java:70)



I tried deleting write.lock. But then I got errors about corrupt search
indexes. So I gave up and scrubbed the search indexes. Which seems to have
shut it up :-).

Anyway, the script now works, but it isnt the cleanup upgrade path I've
seen..



=======================================
Robert Cohen
Systems & Desktop Services
Division of Information
R.G Menzies Building
Building 2


The Australian National University
Canberra ACT 0200 Australia
 
T: +61 2 6125 8389
F: +61 2 6125 7699
http://www.anu.edu.au
 
CRICOS Provider #00120C
=======================================

-------------- next part --------------
A non-text attachment was scrubbed...
Name: content-runconversion.sh.anu
Type: application/octet-stream
Size: 2957 bytes
Desc: not available
Url : http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20100429/9e56ff5d/attachment.obj