Subject: Re: [vserver] Problem with stable 2.2.0.5 and kernel 2.6.22.14
From: Herbert Poetzl <herbert@13thfloor.at>
Date: Tue, 18 Dec 2007 11:43:08 +0100

On Sat, Dec 15, 2007 at 11:19:01AM +0100, Oliver Paulus wrote:
> Hello,
> 
> I am using kernel 2.6.22.14 with the latest stable VServer patch version
> 2.2.0.5. I have similar problems described by "Albert Czarnecki" -
> http://archives.linux-vserver.org/200712/0031.html.
> 
> I have made some tests to check if the problem can be reproduced. I can
> reproduce the problem with one VServer instance which has a squid proxy
> installed. The problem seems to be on a VServer stop command when the VServer
> needs more time to stop. Here is my output:
> 
> Stopping periodic command scheduler: crond.
> Stopping mail-transfer-agent: nullmailer.
> Stopping internet superserver: inetd.
> Stopping OpenBSD Secure Shell server: sshd.
> Saving the system clock..
> Cannot access the Hardware Clock via any known method.
> Use the --debug option to see the details of our search for an access method.
> Stopping Squid HTTP proxy: squid Waiting..................done.
> .
> Stopping system log daemon: syslogd.
> Asking all remaining processes to terminate...done.
> /usr/lib/util-vserver/vserver.stop: line 85: 18657 Getötet            
> "${NICE_CMD[@]}" ${USE_VNAMESPACE:+$_VNAMESPACE --enter "$S_CONTEXT" -- }
> $_VCONTEXT $SILENT_OPT --migrate --chroot --xid "$S_CONTEXT" --
> "${INITCMD_STOP[@]}"
> A timeout occured while waiting for the vserver to finish and it will
> be killed by sending a SIGKILL signal. The following process list
> might be useful for finding out the reason of this behavior:
> 
> ----------------------------------------------------------------------
> 18657    53 vhost3        pts/1    S+     0:00  |       \_ /bin/sh
> /etc/init.d/rc 6
> 18758    53 vhost3        pts/1    S+     0:00  |           \_ /bin/sh
> /etc/rc6.d/S20sendsigs stop
> 18760    53 vhost3        pts/1    S+     0:00  |               \_ sleep 5
> ----------------------------------------------------------------------
> 
> In my /var/log/kern.log I get the following:
> 
> Dec 15 11:18:02 hal9000 kernel: vxW: [xid #0] !!! limit: ffff8100133fe078[VM,9]
> = 31 on exit.
> 
> This happens only with *this* VServer instance. My other VServer
> instances with webservers, mailservers, etc. does not have this
> problem. I have checked the configuration of this VServer - and
> it seems to be correct. 

> Is this some kind of timeout? 

yes

> Can it be configured? 

yes

/etc/vservers/<vserver-name>/apps/vshelper/sync-timeout

 The timeout in seconds which is used when synchronising 
 vserver startup/shutdown with the vshelper. 
 When not set, 30 seconds will be assumed.

http://www.nongnu.org/util-vserver/doc/conf/configuration.html

> The SIGKILL problem was in the previous version of my VServer
> installation too. 

> But the entry in the kernel log seems to be new.

that message looks bogus (context 0 should not be reported at all)
maybe a debian bug, never seen it in mainline (yet)

HTH,
Herbert

> Other info:
>  1. using Debian Etch (4.0)
>  2. VServer utils from Etch (stable)
> 
> Regards
> --
> Oliver Paulus
> 
> OpenPGP
> Key id: 28D9C44F
> Fingerprint: EADA 62FC 07DC 3361 A3D6  4174 2DE3 C027 28D9 C44F
> Public Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x28D9C44F