[Building Sakai] Question About Realm Caching

[Stoufer,R M]

Hi Sam,

Here are the results for anything realm related:

org.sakaiproject.authz.impl.DbAuthzGroupService.realmRoleGroupCache: count:5000 hits:1657342 misses:150730 hit%:91

org.sakaiproject.db.BaseDbFlatStorage.SAKAI_REALM_PROPERTY: count:10000 hits:1416478 misses:385363 hit%:78

[ name = org.sakaiproject.authz.impl.DbAuthzGroupService.realmRoleGroupCache status = STATUS_ALIVE eternal = false overflowToDisk = false maxEntriesLocalHeap = 5000 maxEntriesLocalDisk = 10000000 memoryStoreEvictionPolicy = LRU timeToLiveSeconds = 7200 timeToIdleSeconds = 7200 diskPersistent = false diskExpiryThreadIntervalSeconds = 120 cacheEventListeners:

[ name = org.sakaiproject.db.BaseDbFlatStorage.SAKAI_REALM_PROPERTY status = STATUS_ALIVE eternal = false overflowToDisk = true maxEntriesLocalHeap = 10000 maxEntriesLocalDisk = 10000000 memoryStoreEvictionPolicy = LRU timeToLiveSeconds = 120 timeToIdleSeconds = 120 diskPersistent = false diskExpiryThreadIntervalSeconds = 120 cacheEventListeners:

I just noticed that overflowToDisk is true on the last item.  We have been changing that setting to false most everywhere and I wonder if we should do that here?

Mike


--

Michael Stoufer
Office of Academic Technology
Instructional Computing Services

________________________________
From: Sam Ottenhoff [ottenhoff at longsight.com]
Sent: Wednesday, December 18, 2013 2:00 PM
To: Stoufer,R M
Cc: (sakai-dev at collab.sakaiproject.org)
Subject: Re: [Building Sakai] Question About Realm Caching

Hi Mike,

Maybe your caches are overflowing?  Go to Administration Workspace -> Memory -> Status and share the status results.  The relevant results would be maxed out caches (count=10000) and low hit percentages (good caches would have a hit % greater than 80).

--Sam


On Wed, Dec 18, 2013 at 1:57 PM, Stoufer,R M <kc4uf at ufl.edu<mailto:kc4uf at ufl.edu>> wrote:
We have been looking at Oracle performance on our 2.9.x instance of Sakai and noticed a query that had 1.3 million executions in 3900 seconds.  This was during final exams week with about 14,000 concurrent sessions in progress.  The partial query is:

select 1 FROM SAKAI_REALM_RL_FN srrf, SAKAI_REALM_FUNCTION srf, (select realm_key, role_key from SAKAI_REALM_RL_GR where ACTIVE = '1' and USER_ID = :1 union select -1 as realm_key, -1 as role_key from dual) srrg where rownum = 1 and srrf.realm_key in (select realm_key from SAKAI_REALM.REALM_ID IN....

Unfortunately, that is all I have of the query.

My question is if this is part of the realm caching being done and if not should it be?  It seems that with this many executions in a little more than an hour there may be a more efficient way to handle this.

Thanks for any help!

Mike

--

R. Michael Stoufer
Academic Technology – ICS
University of Florida


_______________________________________________
sakai-dev mailing list
sakai-dev at collab.sakaiproject.org<mailto:sakai-dev at collab.sakaiproject.org>
http://collab.sakaiproject.org/mailman/listinfo/sakai-dev

TO UNSUBSCRIBE: send email to sakai-dev-unsubscribe at collab.sakaiproject.org<mailto:sakai-dev-unsubscribe at collab.sakaiproject.org> with a subject of "unsubscribe"

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20131218/fba4a1b5/attachment.html