Re: [Nbd] oops with timeout option on debian wheezy kernel
- To: Hermann Lauer <Hermann.Lauer@...1489...>
- Cc: nbd-general@lists.sourceforge.net
- Subject: Re: [Nbd] oops with timeout option on debian wheezy kernel
- From: Michal Belczyk <belczyk@...1274...>
- Date: Tue, 11 Feb 2014 23:20:08 +0100
- Message-id: <20140211222008.GE95782@...1273...>
- In-reply-to: <20140210120607.GI19189@...1490...>
- References: <20140210120607.GI19189@...1490...>
On Mon, Feb 10, 2014 at 01:06:07PM +0100, Hermann Lauer wrote:
> Dear nbd kernel maintainers,
>
> according to a mail from 21 July 2011 (http://comments.gmane.org/gmane.linux.drivers.nbd.general/985)
> this is due to the ioctl(nbd, NBD_SET_TIMEOUT, timeout) not working.
>
> Here on debian wheezy (Linux www1 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1+deb7u1 x86_64 GNU/Linux) a disconnection of the server host yields hanging
> of any "cat /proc/mdstat" on the raid1 on top of the nbd device
> and to the appended oops (only two shown) of the kernel. Only
> bringing back the nbd server allows further commands on the md device.
>
> Any news/ideas on that subject ?
No code has been pushed upstream to fix it since then...
You may want to take a look here:
https://www.mail-archive.com/nbd-general@lists.sourceforge.net/msg01513.html
https://www.mail-archive.com/nbd-general@lists.sourceforge.net/msg01549.html
I will post the patch against the most recent tip later this week.
--
Michal Belczyk Sr.
Reply to: