/etc/webmin/server-manager/status.pl: Unknown Xen config line

Hi Jamie,

Ever since you gave me the new Cloudmin image with the bridge support for creation I've been getting sporadic Cron emails about Unknown Xen config line.

I've now gotten 4 of them total, each one different. Two we're on Sept 11th (22:45 & 23:00) and two were Sept 13th (04:50 & 09:50).

From: Cron Daemon <root@xxx>
Subject: Cron <root@xxx> /etc/webmin/server-manager/status.pl
Date: September 11, 2009 10:45:07 PM EDT
To: root@xxx

Unknown Xen config line ty1 clocksource=jiffies"
From: Cron Daemon <root@xxx>
Subject: Cron <root@xxx> /etc/webmin/server-manager/status.pl
Date: September 11, 2009 11:00:25 PM EDT
To: root@xxx

Unknown Xen config line ••••••••••••••••••••••••••••••••••••••••••••
From: Cron Daemon <root@xxx>
Subject: Cron <root@xxx> /etc/webmin/server-manager/status.pl
Date: September 12, 2009 4:50:25 AM EDT
To: root@xxx

Unknown Xen config line 1.139"
From: Cron Daemon <root@xxx>
Subject: Cron <root@xxx> /etc/webmin/server-manager/status.pl
Date: September 12, 2009 9:50:25 PM EDT
To: root@xxx

Unknown Xen config line vncpasswd = "F

Let me know what else I can provide to help debug this.

Status: 
Active

Comments

Could you attach one of the Xen .cfg files (typically under the /xen directory on the host) from a virtual system for which you got this error?

That's the problem, since these emails are coming from my Cloudmin system I don't know which Xen config file is being parsed when the error occurs. However, most of my config files look like this:

kernel = '/xen/vmlinuz-vm2-xenU'
ramdisk = '/xen/initrd.vm2.xenU.img'
memory = 1024
name = 'xen'
address = '1.2.3.4'
netmask = '255.255.255.192'
disk = ['phy:/dev/xen-vg/xen_img,sda1,w']
root = '/dev/sda1 ro'
vif = ['bridge=customer', 'bridge=storage']
extra = 'xencons=tty console=tty1 clocksource=jiffies'
vnc = 5905
vnclisten = "1.2.3.4"
vncdisplay = 5
vncpasswd = "******"

Do any of your config files have clocksource=jiffies in them?

Also, does this email get sent continually, or just occasionally?

Yes, I was dealing with a clock skew issue on a couple of machines and some of the config files have clocksource=jiffies (Turned out it wasn't the issue but I haven't had a reason to remove it yet).

The email is occasionally. It's almost like the data stream get's corrupted while Cloudmin is trying to read the config of a remote machine. Sometimes it contains actual text from my config files, other times it's non-printable or extended characters (got 2 with euro's today).

Just got a real interesting one

From: root@xxx (Cron Daemon)
To: root@xxx
Subject: Cron <root@xxx> /etc/webmin/server-manager/status.pl
Date: Wed, 16 Sep 2009 11:30:26 -0400 (EDT)

SIGPIPE handler "ignore" not defined.

That SIGPIPE error looks un-related, but harmless ..

Coming back to the original issue, I suspect you are correct that the Xen .cfg files are not being fetched properly from the Xen host system. Do you see any SSH-related errors in /var/log/messages on the host system?

Just went through all my logs and couldn't find any ssh errors around the times of the cron emails.

Are you still getting these emails, or did they only happen once?

This is tough to debug, as it looks like an scp of those files was only partially complete a few times..