All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peri Hankey <mpah@thegreen.co.uk>
To: "Jérôme Petazzoni" <jp@enix.org>
Cc: xen-devel@lists.sourceforge.net
Subject: Re: Problems with dom0 and low memory
Date: Tue, 07 Dec 2004 12:51:31 +0000	[thread overview]
Message-ID: <41B5A753.8030708@thegreen.co.uk> (raw)
In-Reply-To: <41B588B5.9010309@enix.org>

I'm not sure whether this was related to low memory, but I have seen an 
ext3 journal abort like this on a xenU domain, and have also had one 
occurrence of ext3 file corruption on the same domain. Not while I was 
active on that  domain.

It seems that a worm was probing for one of the IIS buffer overflow 
exploits: initially I saw a lot of "\x90\x90\x90\x90\x90\x90\x90\x90..." 
sequences in /lib/modules/2.6.9-xenU/modules.alias, which didn't go down 
very well with the boot process. I fsck'ed the beast, letting it unpick 
quite a tangle. Subsequently I have seen large SEARCH blocks containing 
that (and similar garbage) appear in the httpd/access_log, and a bit 
googling about suggested that it was probably one of the IIS worms.  
There's also an attempt to POST a minegrew$ dll to the system - also a 
known exploit.

I haven't been able to pin any of this down more closely, and am a bit 
concerned about the ext3 file system corruption.

Regards
Peri

Jérôme Petazzoni wrote:

>
> I am running a Xen box where I need a lot of memory for a given domain 
> (which runs memory-hungry java apps). I had reduced the memory of 
> domain0 to 42000KB. It *seems* that after running OK for a while, a 
> low memory condition on domain0 messed during a short time the VBD of 
> the other domains, and they got a burst if I/O errors. After this 
> event, the other domains were still running, and their VBD were still 
> working, but the burst of I/O errors aborted the ext3 journal on the 
> filesystems (causing them to go read-only).
>
> Here are the kernel logs of both domains if you want to check that 
> it's indeed a low memory condition problem and not something else 
> (which would then be a "real" bug).
>
> I increased the memory allocated to domain0, but it would be 
> interesting to know if there is a way to avoid those problems (I 
> thought about increasing /proc/sys/vm/min_free_kbytes ; would it be 
> useful here?)
>
> The dom0 and domU kernels, as well as xen itself, the associated 
> symbols, and the same log files are located here, if someone needs them:
> http://skaya.enix.org/webs/xen/dom0-lowmem/
>


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

  reply	other threads:[~2004-12-07 12:51 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-12-07 10:40 Problems with dom0 and low memory Jérôme Petazzoni
2004-12-07 12:51 ` Peri Hankey [this message]
2004-12-07 17:35 ` Keir Fraser
2004-12-07 19:47   ` Derek Glidden
2004-12-07 21:47   ` Jérôme Petazzoni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=41B5A753.8030708@thegreen.co.uk \
    --to=mpah@thegreen.co.uk \
    --cc=jp@enix.org \
    --cc=xen-devel@lists.sourceforge.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.