Hi,
I had some problems with CacheManager resizing efficiency
(see https://issues.apache.org/jira/browse/JCR-1112).
Recently I exposed public method on CacheManager which allows to setup
resize interval programatically. By default CacheManager is doing
recalculation every second, and it seems that this method is time
consuming especially when number of managed caches grows.
Thanks, Przemo
Hi all,
I have a problem regarding CacheManager size under heavy load and system
slowdown.
CacheManager start with:
16:28:13,476 INFO CacheManager:125 - resizeAll size=78
after it grows too much:
17:28:23,079 INFO CacheManager:125 - resizeAll size=122
17:28:24,143 INFO CacheManager:125 - resizeAll size=224
17:28:25,270 INFO CacheManager:125 - resizeAll size=382
17:33:24,567 INFO CacheManager:125 - resizeAll size=468
...
19:09:35,674 INFO CacheManager:125 - resizeAll size=162467
19:11:03,601 INFO CacheManager:125 - resizeAll size=162737
19:12:33,840 INFO CacheManager:125 - resizeAll size=163095
Then system starts slowdown and some errors occured.
I use a Session Pool to share some session to 'read-only' users. Creation of
the pool that starts with webapp and sessions are never logged only (only in
case of errors and recreated) to be shared among users. I share 15 sessions
for each of the 2 workspaces I use.
Please can anyone address me to an analysys of this problem?
I use JR 1.3.1 patched with JIRA-905 patch to avoid duplicated entries in
index searches.
Server is Oracle 10g on Solaris. Db used is also Oracle10g.
Another point, I have a big collection of blobs in db (but I don't thinks it
could be the problem).
It's pretty urgent.
Thanks a lot in advance.
BR,
Andre.
