How can I increase verbosity of ejabberd logs ?
I asked this becouse I have self-written jabber bot on python, and in some cases (when it send single, usual, not big message to any user through ejabberd) it makes ejabberd to eat up to 700 Mb of memory and more (I have 320M RAM and 500M swap), until OS do - "kernel: pid 499 (beam), uid 522, was killed: out of swap space". At this moments i cannot stop it with # ejabberdctl ejabberd@`hostname -s` stop (it says smth about that node is down) and only killall -u ejabberd helps.
My guess - that this is becouse of Mnesia - I noticed similar behavior while
# ejabberdctl delete-expired-messages
and while start after unclean shutdown (killall).
In normal state (with users, but without bot) ejabberd eat's about 80M. And I did not notice such problems with normal users, only with bot :)
I cannot (yet :)) find a bug in my bot, so I simply want to know what ejabberd do at this moments.
P.S. ejabberd-1.0.0, erlang-r10b9,1 at FreeBSD 5.4-R.
Thanks for any help!
Hmmmm...This situation can
Hmmmm...https://xxx:5280/admin/server/xxx/user/dds/queue/ the beam proccees started to eat more than 500M (and keep growing) of memory for a long time (for now I'm waiting about 20 min.).
This situation can be repeated on my server with looking properties (through web interface) of user which have 4690 offline messages. After typing at browser
What can I do in this situation?
Is there a way to delete offline messages for all users which have more then 100 (for example) offline messages? or to keep only 100 last (by date) messages?
Is there a way to delete users which never log on to server? or did not logon within last year?
Thanks!
Re: Hmmmm...This situation can
Re: verbosity of ejabberd logs
How can I increase verbosity of ejabberd logs ?
I've just added this page to the ejabberd book:Increase log verbosity . If it doesn't work, tell me.
I hope this helps you to find the problem you are experiencing.