APT_BadAlloc: Heap allocation failed.

rajeevn80 · Post by **rajeevn80** » Sun Nov 27, 2005 11:11 pm

hi guys,
when i run a px job the job abborts with the following message : -

APT_CombinedOperatorController(2),0: Caught exception from runLocally(): APT_BadAlloc: Heap allocation failed.

can anyone help me on this...is it related to some memory issue...???

ArndW · Post by **ArndW** » Mon Nov 28, 2005 2:05 am

Have you checked disk space during execution of the job? If the error occurs after a bit of time then this might be the cause, if it happens right at the outset then you might have access issues.

rajeevn80 · Post by **rajeevn80** » Mon Nov 28, 2005 3:07 am

the job runs for 5 mins and then aborts.... this job uses 5 lookups and the lookup files are very huge.... hoping that this could be a memory problem, i changed all the lookups to joins(inner joins)... but still the job aborts with the same message...

since no stage names are mentioned in the director log... i assumed that this could be the memory issue...

any solutions ..??

ArndW · Post by **ArndW** » Mon Nov 28, 2005 4:02 am

Rajeev,

the lookups will use up temporary disk space. If you use a temporary 1-node configuration file does the problem go away?

ray.wurlod · Post by **ray.wurlod** » Mon Nov 28, 2005 5:19 am

"Heap" always refers to memory, but it's virtual memory so can be exhausted by running out of spill space (scratch disk resource) or of swap space, depending on what's actually occurring. So I think you are on the right track suspecting memory, but probably need to do some more detective work.

There are some environment variables that can cause monitoring statistics to be captured. These may help. You might also compile in debug mode, though I don't believe this gives you memory usage statistics. And, of course, you can monitor from the operating system level, using commands such as vmstat.

Ultramundane · Post by **Ultramundane** » Tue Nov 29, 2005 10:09 am

I had this same issue on AIX. I had to increase the amout of memory that the osh program could allocate. I modified it so that it could allocate 2 gb instead of the default 512 MB.

cd $APT_ORCHHOME/bin
/usr/ccs/bin/ldedit -bmaxdata:0x80000000/dsa osh

ak77 · Post by **ak77** » Tue Nov 29, 2005 11:48 am

I had similar kind of error
But something strange was goin in the Server/Database
So the Admin just restarted asking all of us to get out of the box and it all worked fine after that

Talking to the DBA might help

Kishan

pavankvk · Post by **pavankvk** » Thu Dec 08, 2005 3:25 am

How can we know our current allocation for osh?

pavankvk · Post by **pavankvk** » Thu Dec 08, 2005 3:26 am

i get the following messages in the log..

APT_CombinedOperatorController(1),0: The current soft limit on the data segment (heap) size (2147483645) is less than the hard limit (2147483647), consider increasing the heap size limit

APT_CombinedOperatorController(1),0: Current heap size: 2,141,571,120 bytes in 580,463 blocks

I think i have 2 gigs..am i correct?