is <https://luceeserver.atlassian.net/browse/LDEV-...
# lucees
steveduke
04/06/2022, 5:28 PMis https://luceeserver.atlassian.net/browse/LDEV-3924 fixed in RC3 ?
b
bdw429s
04/06/2022, 5:30 PM@steveduke The ticket is still in the backlog and unresolved, so it would not appear ot be fixed in any version of Lucee.
z
zackster
04/06/2022, 7:20 PMI'm trying to repro at the moment
👍 1
zackster
04/06/2022, 7:48 PMthis is my first test, with idle timeout set to 1m for the connection
zackster
04/06/2022, 7:50 PMobviously more than times="10"
zackster
04/06/2022, 7:50 PMinterestingly, at the start mysqld is 60% cpu and lucee 10%... after a while lucee goes to 70%
zackster
04/06/2022, 8:32 PMbreaks out the debugger
zackster
04/06/2022, 8:34 PM92k pagecontexts seems high too, i had 4 requests going
zackster
04/06/2022, 8:35 PMmy poor old laptop, running at 100%
b
bdw429s
04/06/2022, 8:45 PMYour async array each is creating a new pc for each iteration
bdw429s
04/06/2022, 8:45 PMI don't even think Lucee pools those
bdw429s
04/06/2022, 8:48 PMI pointed out in LDEV-2559 that Lucee didn't need to create more pc than it had threads in the pool (default is 20) but I don't think Micha ever changed that which is really too bad
bdw429s
04/06/2022, 8:48 PMpcs should be reset and re-used, not created by the thousands
z
zackster
04/07/2022, 6:48 PMAnyway, we can't reproduce so far, @micha is going to have a look tomorrow. It's not a blocker for the first stable release but we're trying. We might end up backporting from 6.0 the Apache commons pooling, but that would land in 5.3.10
👍 3
3 Views