[systemd-devel] System stability when journald locks up

Marti Raudsepp marti at juffo.org
Wed May 30 12:36:21 PDT 2012


On Wed, May 30, 2012 at 4:45 PM, Lennart Poettering
<lennart at poettering.net> wrote:
> OK, there indeed was a loop here, where we ended up processing messages
> we ourselves wrote to kmsg. I have fixed that now in git. Could you test
> please if this solves your problems?

Thanks. Yes, looks like this fixes my issue.

I filled my disk, made sure that journald starts consuming 100% CPU.
Then replaced /usr/lib/systemd/systemd-journald with one built from
git, killed and the problem disappeared.

On Tue, May 29, 2012 at 11:29 PM, Lennart Poettering
<lennart at poettering.net> wrote:
> The journal is still very new. I think so far it is quite
> stable, but there is definitely more work necessary to make it rock
> solid in all corner cases.

Well, any concrete ideas? Locking the user out of his/her own system
is the best way to become hated by sysadmins. I certainly don't want
to see journald on my servers until that's addressed.

Maybe a client-side timeout in libsystemd-journal? While that could
still effectively crash applications by slowing them down too much, at
least it's possible to log in to inspect and fix the issue.

Regards,
Marti


More information about the systemd-devel mailing list