[
https://issues.apache.org/jira/browse/DERBY-1781?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12535437
]
Thomas Nielsen commented on DERBY-1781:
---------------------------------------
Bryan, thanks for shedding some light on the DRDAConnThread lifespan. I was
expecting them to stay around, but I also was expecting them to be reused the
second time I ran the repro script. This does not seem to be the case, as I see
new DRDAConnThreads being created, and the old ones stay suspended. I guess
there is a pool or a queue that these are placed into, but are they only reused
after N threads are created?
The majority of surviving generations are StoredRecordHeader and RecordId
objects as previously stated. Of the StoredRecordHeader objects 52% originate
from store.raw.data.StoredPage.recordHeaderOnDemand() and 47% from
store.raw.data.StoredPage.storeRecordForInsert().
There's about 2200 live StoredRecordHeader objects consuming just above 100K of
memory. This is probably touching on the work Knut is currently doing in
DERBY-3130.
The vast majority of RecordId objects are part of a StoredRecordHeader object,
and originate from calls to StoredRecordHeader.getHandle().
The number of surviving generations is normally a measure of probable memory
leaks, but in our case it does not seem to be the case at first sight. The
StoredRecordHeader objects are referenced from the StoredPage objects in the
page cache. They are not freed until they are evicted or cleared from the
cache. They are intentionally kept from GC and will add to the number of
surviving generations. The pages are not cleared from the cache until the cache
is full and more pages need loading. The new pages will then have references to
new StoredRecordHeader object and the unused references will be GC'ed.
This investigation was done on the main trunk. I'll give 10.2.1.1 a spin as
well to validate that there are no significant changes here from 10.2.1.1 to
trunk.
> Process handles appear to be leaking in queries using an IN clause during
> concurrent DB access
> ----------------------------------------------------------------------------------------------
>
> Key: DERBY-1781
> URL: https://issues.apache.org/jira/browse/DERBY-1781
> Project: Derby
> Issue Type: Bug
> Components: SQL
> Affects Versions: 10.1.3.1
> Environment: Windows XP, Java 1.5.0_05
> Reporter: Mark Hellkamp
> Attachments: SqlStressTest.java
>
>
> We are currently using Derby embedded in our web application running on
> Windows. When processing multiple concurrent requests we have noticed that
> the Java process handle count continues to increase until the machine becomes
> unresponsive. I was able to isolate the problem to Derby by running the
> database in network mode in another process. Further investigation showed
> that the problem could be reproduced using a select statement that has an IN
> clause with multiple entries on the primary key column. Spawning multiple
> threads running the same query causes the handle count to increase
> considerably on the Derby process. The problem occurs in version 10.1.3.1 and
> 10.2.1.1 (even worse) in both embedded and network mode. The attached test
> program duplicates the problem. Start Derby in network mode (using
> startNetworkServer.bat) and run the enclosed test program. The handle count
> on the Derby process will increase and never go down short of restarting
> Derby. Using 10.2.1.1 the handle count for the Derby process goes somewhere
> between 1400-1500 with just two threads in my environment.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
