[Deploying Sakai] Swapping behavior on 8gb machine with 6gb JVM heap

[Martin B. Smith]

Good morning all,

I know a lot of folks out there are using app servers with 8 gigs of RAM 
and limiting the JVM to 6 gigs of heap. Are any of you also seeing any 
of your JVM process put into swap?

I'm noticing that a couple hundred megs of swap get used up after our 
app tier is running for a few days. I suspect that NFS activity causes 
FS cache to be slightly favored over the JVM after a while running.

Is anyone monitoring their swap? Are you just turning it off? Are you 
turning down other JVM memory parameters so that the sum of all memory 
parameters in your JVM reach 6gb (for instance, a 5gb heap and 512mb of 
permgen and 512mb of newgen)? Perhaps you have the vm.swapiness kernel 
parameter set?

This isn't really hurting us, but it seems like it might be better to 
avoid JVM contents in swap. I did the obligatory search in mailing list 
archives, but didn't come across anything that specific to my question.

Thanks in advance,
-- 
Martin B. Smith
smithmb at ufl.edu - (352) 273-1374
CNS/Open Systems Group
University of Florida

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5497 bytes
Desc: S/MIME Cryptographic Signature
Url : http://collab.sakaiproject.org/pipermail/production/attachments/20100914/bb23b393/attachment.bin