[prev in list] [next in list] [prev in thread] [next in thread] 

List:       osiris
Subject:    [osiris] Re: scheduler going dead?
From:       "Alexei_Roudnev" <Alexei_Roudnev () exigengroup ! com>
Date:       2005-12-13 2:27:52
Message-ID: 123f01c5ff8c$d14791c0$6f31a8c0 () sjc ! exigengroup ! com
[Download RAW message or body]

I have the same in the very very old osirisd (version 2), running on
Windows. (I configured one of hosts to send messages every time, even when
system do not find any changes, so I can see when scans stopped easily).

One problem rely to the dead conenctions, which (sometimes) heaps up until
they block normal one. Other looks as scheduler bug - sometimes scheduler
skips scans.

I cannot confirm it on the new version because I (still) did not upgraded
(clients are incompatible), so may be this is another issue.
Anyway, I prefer to have embedded machanism which protects us from such
behavior (may be, by restarting osirisd if it see a problems).

----- Original Message ----- 
From: "David Thiel" <lx@redundancy.redundancy.org>
To: <osiris@lists.shmoo.com>
Sent: Monday, December 12, 2005 5:55 PM
Subject: [osiris] Re: scheduler going dead?


> On Mon, Dec 05, 2005 at 03:08:33PM -0500, David Vasil wrote:
> > David Thiel wrote:
> > > On a system that scans several hundred hosts, I've observed that
> > > after a while, scans just stop happening, even though osirismd seems
> > > to respond just fine. A typical situation is that after 3 or 4 days
> > > of running normally, the logs are empty, and no scans have happened -
but
> > > when I log into the management console, log entries are generated.
Killing
> > > the osirismd processes(and I do mean killing, as some just won't go
away
> > > with a regular SIGTERM) and restarting solves the problem.
>
> Ok, I've had this problem again. This is a fbsd 6.0-RELEASE box. I
> looked at the processes running:
>
> | |-+= 23807 osiris /usr/local/sbin/osirismd -r /usr/local/osiris
> | | \-+- 23808 osiris /usr/local/sbin/osirismd -r /usr/local/osiris
> | |   |--- 23809 osiris [osirismd]
>
> (about 900 more osirismd procs)
>
> | \-+- 00582 root /usr/local/sbin/osirisd -r /usr/local/osiris
> |   \--- 00583 osiris /usr/local/sbin/osirisd -r /usr/local/osiris
>
> A ktrace -dip on 23807 shows that it's spawning processes to handle
> incoming admin requests from the CLI, but doing nothing else. None of
> the child procs are doing anything at all. The state for the hung procs
> is "RE", so they're supposedly trying to exit. Procs 23807, 23808, and
> 583 are idle.
>
> Wish I had more info, but all the processes are literally doing nothing
> unless a console is spawned. What now?
> _______________________________________________
> osiris mailing list
> osiris@lists.shmoo.com
> https://lists.shmoo.com/mailman/listinfo/osiris
>

_______________________________________________
osiris mailing list
osiris@lists.shmoo.com
https://lists.shmoo.com/mailman/listinfo/osiris
[prev in list] [next in list] [prev in thread] [next in thread]