Perl 6 - the future is here, just unevenly distributed

IRC log for #openam, 2015-07-13

| Channels | #openam index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:18 MegaMatt joined #openam
01:29 MegaMatt joined #openam
05:27 ramteid joined #openam
06:43 SteveFerris joined #openam
06:49 tsmalmbe joined #openam
07:35 SteveFerris_ joined #openam
07:40 aldaris joined #openam
08:16 balo joined #openam
08:30 aldaris Good morning
08:37 jjpp good morning.
08:37 balo aldaris: https://twitter.com/majorpet​ya/status/620510922130477057 you missed the context here :P
08:38 aldaris hogyhogy?
08:39 balo that you release openam. i mean i know it's about openam but i'm not sure if everyone knows (and they should know aobut it :)
08:40 aldaris I don't care about that :)
08:43 aldaris1 joined #openam
08:43 aldaris1 ah, balo, you've tricked me :)
08:45 aldaris it's quite exhausting to build security patches for AM :(
09:04 balo i can imagine
09:08 KermitTheFragger joined #openam
09:27 aldaris joined #openam
11:29 MegaMatt joined #openam
11:38 aldaris joined #openam
12:06 mckeanbs joined #openam
12:49 aldaris nobody has OpenAM issues nowadays? :)
12:49 MegaMatt Product too perfect.
13:05 jjpp aldaris: well.. yes I do have a problem. but it is way too vague at the moment to report it. :)
13:12 jjpp (the problem is that occasionaly active session count will continue to grow. until we restart relevant node or the configured limit is reached)
13:13 jjpp we got it to be ~22000 today while there was just ~3100 entries in CTS repository
13:13 MegaMatt what kind of sessions are they?
13:14 MegaMatt (I bet they are 3.5.0 agent sessions?)
13:14 jjpp the thing is that as far as i remember from previous times -- these seem to be completely phantom -- there is no real sessions.
13:15 aldaris https://bugster.forgerock.o​rg/jira/browse/OPENAM-5835
13:15 MegaMatt You can look at the session objects in the JVM heap
13:15 MegaMatt Aldaris stole my thunder ;)
13:15 aldaris we missed endless session loops so much that we reintroduced them accidentally :)
13:15 jjpp :)
13:15 MegaMatt If you have a 3.5.0 agent, it’s probably that bug
13:15 aldaris I think there is a possibility that even the servers could do that, but I'm not 100% sure about that
13:16 MegaMatt So far I’ve only seen it (consistently) with 3.5.0 agent
13:16 jjpp I would say that we do not have any real agents deployed. but then again i have been away and a lot of weird things has happened in mean time. :)
13:16 aldaris then anything that uses 12 clientsdk?
13:16 MegaMatt So, when your session count goes high, take a heap dump
13:16 MegaMatt and take a look into the session
13:16 jjpp and it used to happen even when we did not have agents for sure.
13:16 MegaMatt Will give you a clue as to what is creating them
13:17 jjpp otoh, we have 2 daui nodes
13:17 jjpp yeah, scripting the memory dump is one thing i plan to do in near future.
13:18 MegaMatt Fire up a heap analyzer and run an OQL like SELECT clientID FROM com.iplanet.dpro.session.Session
13:18 jjpp shouldn't session list with ssoadm do the same?
13:18 jjpp essentially
13:18 MegaMatt I don’t think that shows agent sessions
13:19 jjpp perhaps it had an option for that. but yes, full heap dump is probably good idea anyway.
13:20 jjpp i think there used to be some cases where some clean-up threads did die, prehistorically. and that was the reason why session counter was not decreased. that should be visible from the dump as well.
13:20 MegaMatt If it shows all the sessions, including agent sessions .. then sure.. but I’m guessing it will only show user sessions - just like the control panel
13:22 jjpp well, i'm not even sure what it is what it does show. i kind of remember list-sessions showing a lot less sessions than there was in token repository (even for both of our nodes added together). so.. it might have some filter.
15:33 pcypher joined #openam
15:37 asyd joined #openam
15:37 jjpp joined #openam
15:37 ikonia joined #openam
15:37 GrosSeb joined #openam
16:29 pcypher joined #openam
17:01 pcypher_ joined #openam
17:35 pcypher joined #openam
17:52 pcypher joined #openam
19:14 aldaris joined #openam
19:27 aldaris joined #openam
19:29 pcypher joined #openam
19:31 aldaris joined #openam
19:44 aldaris joined #openam
19:49 aldaris joined #openam
20:00 pcypher joined #openam
20:10 aldaris joined #openam
20:23 auke- joined #openam
20:31 aldaris joined #openam
21:00 pcypher joined #openam
21:23 aldaris joined #openam
21:27 aldaris joined #openam
21:52 aldaris joined #openam
22:24 pcypher joined #openam
22:26 pcypher_ joined #openam
22:39 MegaMatt joined #openam

| Channels | #openam index | Today | | Search | Google Search | Plain-Text | summary