Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 1 Apr 2011 18:20:13 +0200 (CEST)
From:      Oliver Fromme <olli@lurza.secnetix.de>
To:        freebsd-geom@FreeBSD.ORG
Subject:   gmirror question
Message-ID:  <201104011620.p31GKDbM036600@lurza.secnetix.de>

next in thread | raw e-mail | index | archive | help
Hi,

I just logged into a machine when my zsh shell died before
even printing the prompt.  The machine runs with a gmirror
of two disks (ad0 and ad1).  ad1 has always been a little
flaky, throwing errors every few weeks or months, then being
disconnected by gmirror.  This never caused any problems in
the past; the machine just continued running with ad0 alone.
A reboot always brought ad1 back, and after syncing gmirror,
everything continued running fine for some weeks or months.

However, when the same thing happened today, gmirror wasn't
able to reroute all requests to ad0, it seems.  That's why
my shell died.  I would like to understand *why* this could
happen.  I thought that it's the purpose of gmirror to keep
everything running when one disk dies.

This is what the kernel logged during the event:

17:50:42 ad1: TIMEOUT - READ_DMA retrying (1 retry left) LBA=778951
17:50:53 ad1: TIMEOUT - READ_DMA retrying (1 retry left) LBA=778215
17:51:36 unknown: TIMEOUT - READ_DMA retrying (1 retry left) LBA=777767
17:51:36 ata3: timeout waiting to issue command
17:51:36 ata3: error issuing READ_DMA command
17:51:36 GEOM_MIRROR: Request failed (error=6). ad1[WRITE(offset=155136, length=16384)]
17:51:36 GEOM_MIRROR: Device gm0: provider ad1 disconnected.
17:51:36 g_vfs_done():mirror/gm0s1a[READ(offset=398176256, length=45056)]error = 5
17:51:36 vnode_pager_getpages: I/O read error
17:51:36 vm_fault: pager read error, pid 78331 (zsh)
17:51:36 pid 78331 (zsh), uid 1000: exited on signal 11 (core dumped)

In particular, I wonder about the third line which mentions
an "unknown" device.  I assume this comes from ad1 while it
is being disconnected.

Can anybody explain why the I/O error on gm0s1a happened?

If it matters, this is a single-core machine (not SMP)
running FreeBSD running 8-stable csupped on 2010-10-22.

Best regards
   Oliver

-- 
Oliver Fromme, secnetix GmbH & Co. KG, Marktplatz 29, 85567 Grafing b. M.
Handelsregister: Registergericht Muenchen, HRA 74606,  Geschäftsfuehrung:
secnetix Verwaltungsgesellsch. mbH, Handelsregister: Registergericht Mün-
chen, HRB 125758,  Geschäftsführer: Maik Bachmann, Olaf Erb, Ralf Gebhart

FreeBSD-Dienstleistungen, -Produkte und mehr:  http://www.secnetix.de/bsd

C++: "an octopus made by nailing extra legs onto a dog"
        -- Steve Taylor, 1998



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201104011620.p31GKDbM036600>