Unsolved
This post is more than 5 years old
5 Posts
0
3674
August 5th, 2009 06:00
FP_OUT_OF_MEMORY_ERR in getRetentionClassContext()
When trying to get an FPRetentionClassContext (in Java), I get an out of memory exception, the problem manifests itself in the GetClusterInfo.java example program as well as in a small test, which in effect only does:
myPool = new FPPool(poolAddress);
FPRetentionClassContext retentionClassList = myPool.getRetentionClassContext();
Result is something like:
Exception: System ran out of memory
ClusterCloud::GetRetentionClassContext(-1,FPRetentionClassContext&,UpdateFlag) Context(-)
com.filepool.fplibrary.FPLibraryException: System ran out of memory
at com.filepool.fplibrary.FPPool.getRetentionClassContext(Unknown Source)
at PoolBonk.main(PoolBonk.java:25)
Or FP_OUT_OF_MEMORY_ERR(-10206) in GetClusterInfo
I'm using SDK 3.2.687 on Windows XP with Java 1.5.0_11 against our client's Centera running CenteraStar 4.0.2-3195-1019-20652
Any thoughts why this might be happening here?
Cheers,
-daniel
0 events found
gstuartemc
2 Intern
•
417 Posts
1
August 5th, 2009 07:00
Hi Daniel - does the cluster in question have a large number of Retention Classes defined?
If you run the JCASScript tool against that cluster, are you able to retrieve the Rentetion Classes successfully?
cluesch
5 Posts
0
August 5th, 2009 08:00
Well it doesn't look like I've got too many, but not enough retention classes:
CASScript>rc
No retention classes exist on the primary cluster
CASScript>
So I'm guessing that somewhere in the API someone takes an invalid value for the number of classes (say -1) and uses that for malloc purposes, and since malloc takes unsigned numbers...
But that's just me guessing. How JCASScript can get the retention classes, when the API can't is somewhat fascinating.
Cheers,
-daniel
cluesch
5 Posts
0
August 5th, 2009 09:00
Actually the answer to that is quite simple.
JCASScript is built against (and shipping with) SDK version 3.2.607, and I'm using 3.2.687 (3.2 Patch 4). If I'm running the programs against JCASScript.jar and its DLLs everything works.
So someone broke the SDK (Same issue also on Solaris 10) when fixing other things. So we'd need another Patch, or the previous (3.2.607) version for downloading.
Problem solved, or at least identified.
gstuartemc
2 Intern
•
417 Posts
0
August 9th, 2009 08:00
Thanks Daniel - I have raised a TestRack defect report and this should be fixed in the next patch release.
As a workaround, please create a Retention Class on the cluster in question.