=== Cidan is now known as zz_Cidan | ||
l3on | Hi All.. I'm trying to run bash via upstart and prevent users to have a max num of process children in that session .. is it possible you think ? | 14:47 |
---|---|---|
xSmurf | l3on: sounds like you want cgroups | 15:34 |
l3on | that's what I'm looking at right now ... xSmurf | 15:37 |
l3on | but I can't find a max.tasks settings | 15:37 |
xSmurf | yeah now that I think of it I don't think there is such a setting | 15:44 |
xSmurf | because it could be dangerous as the kernel can't tell whether a process is actually critical or not, CPU limiting is not good enough for you? | 15:44 |
l3on | So, since I want prevent forkbomb .. which limits of cpus and memory you think can be used ? | 15:45 |
xSmurf | I see | 15:45 |
xSmurf | you want a per user ulimit I guess | 15:45 |
l3on | It's odd .. my users connect to my docker with same username | 15:46 |
l3on | so I can't have it by user, I have to create a cgroup at each connection | 15:46 |
l3on | I would like to understand what limits can I use .. | 15:46 |
xSmurf | if they are all sharing the same user then I don't know | 15:47 |
l3on | they are | 15:47 |
xSmurf | if they had their own accounts then ulimit would work | 15:47 |
l3on | I know .. just my architectures does not work with ulimit | 15:48 |
xSmurf | why can't they all have their own user and share a group instead? | 15:48 |
l3on | dunno, not my choice | 15:48 |
xSmurf | I see | 15:48 |
xSmurf | in that case I have no idea, sorry | 15:48 |
=== zz_Cidan is now known as Cidan | ||
=== Cidan is now known as zz_Cidan | ||
=== zz_Cidan is now known as Cidan | ||
mustmodify | Hey, I've been using upstart in trivial ways for a while. I'm looking at using it to manage resque *and* my www service... is there an easy way to have `sudo start app` start both `app-resque` and `app-www` ? | 21:37 |
=== Cidan is now known as zz_Cidan | ||
=== zz_Cidan is now known as Cidan |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!