From: sysadmin_at_qlusters.com
Date: Mon 17 Nov 2003 - 15:48:22 GMT
('binary' encoding is not supported, stored as-is)
('binary' encoding is not supported, stored as-is)
The e-mail you have sent to this domain (qlusters.com) did not reach its indended destination.
Possible reasons: the e-mail address entered is mispelled or the user is none existant.
Note:
---------
You may see the original e-mail below. If you feel that this is error message is incorrect or
you have a question then feel free to press "reply" to this e-mail and e-mail the Administrator.
Thank you,
Qlusters INC.
attached mail follows:
On Mon, Nov 17, 2003 at 06:23:00AM -0500, Matt Ayres wrote:
> I have 2 servers who keep crashing, sometimes every few hours... I am
> still working on a console solution so I don't know the exact place
> where the problem is.
Hi Matt!
> The servers are in a "half hung" state, they respond to PING and will
> open TCP connections, but they won't send any data back.
>
> Is this how yours responds or is it a hard lock?
>
> I've tried with 2.4.22-c17e, 2.4.23pre8-c17h and 2.4.22-vs1.0.0 all with
> the same results.
this is/sounds bad, so we should try to narrow it
down as soon as possible ... the following things
I would suggest, to hunt this down ...
- make some tests with the hardware
(memtest86, temperature, cpu-stability)
- search the log for the following strings
'APIC', 'NMI', 'ACPI', 'oops', 'panic'
- enable the nmi watchdog and make sure it
is active (/proc/interrupts)
- enable magic sysreq, so you can force
a stack trace for all running programms
- look for ERR or MIS counts on a running
machine in /proc/interrupts
- you mention 2.4.22-vs1.00, did you add
any other patches? if so, please try to
reproduce with plain 2.4.22-vs1.00 ...
- if this happens with plain 2.4.22-vs1.00,
please try to reproduce with vanilla
2.4.22 (I know this his hard on a vserver
host, but we have to know, it won't help
to search in vserver code for a 2.4.22
bug/issue ...
- when the machine is just hung, try to
activate the magic sysreq, and dump all
registers as well as the stack trace for
the running tasks ...
HTH,
Herbert
> On Mon, 2003-11-17 at 06:00, James MacLean wrote:
> > Hi folks,
> >
> > 2 nights in a row at 4:00am when netstat runs, the system is Oopsing. I
> > was running vs1.1.2 and am now upgrading to vs1.1.3 but do not see
> > anything that suggests a fix in in the diffs.
> >
> > Possibly I am the only lucky one :)? I remember this happening some time
> > ago in earlier vservers, and someone new the fix. The Oops was within
> > tcp_get_info if I am reading the System.map correctly.
> >
> > JES
> --
> Matt Ayres <matta_at_tektonic.net>
> TekTonic
>
> _______________________________________________
> Vserver mailing list
> Vserver_at_list.linux-vserver.org
> http://list.linux-vserver.org/mailman/listinfo/vserver
_______________________________________________
Vserver mailing list
Vserver_at_list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver
_______________________________________________
Vserver mailing list
Vserver_at_list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver