[Top] [All Lists]

Re: XFS for postgres databases?

To: Eric Sandeen <sandeen@xxxxxxx>
Subject: Re: XFS for postgres databases?
From: marat <marat@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 31 May 2004 13:43:04 +0200
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <Pine.LNX.4.44.0405210956450.9057-100000@xxxxxxxxxxxxxxxxxxxxxx>
Organization: westerngeco
References: <Pine.LNX.4.44.0405210956450.9057-100000@xxxxxxxxxxxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: KMail/1.4.1
Hi Eric ,
Please, see my answer below :

> > I have got some strange problem on Linux RH9 system, kernel 2.4.25
> a kernel.org kernel?

> > The XFS filesystem crashed with message :
> >
> >       Filesystem "sd(8,33)": XFS internal error xfs_btree_check_lblock at
> > line 222 of file xfs_btree.c.  Caller 0xf89b695f f65ada4c f89b7ffa
> > f8a14e7a 00000001 f6e78400 f8a14e6e 000000de f89b695f >             
> > ea6a6dcc e68ffe40 db68dd84 db68dd84 f89b8da5 d3a0d7a8 e3a8d1a0 5f0fbac6
> > 0a05bb82 d87da000 db68dd84 00000000 00000000 f89b695f db68dd84 d87da000
> > Call Trace:    [<f89b7ffa>] [<f8a14e7a>] [<f8a14e6e>] [<f89b695f>]
> > [<f89b8da5>]
> please run this through ksyumoops to get full backtrace information.
Here it is:
Warning (Oops_read): Code line not seen, dumping what data is available

Trace; f89b82a5 <[xfs]xfs_btree_check_sblock+85/100>
Trace; f8a14eab <[xfs].rodata.end+2e0/140b5>
Trace; f8a14e6e <[xfs].rodata.end+2a3/140b5>
Trace; f899dcfc <[xfs]xfs_alloc_lookup+12c/3c0>
Trace; f899dcfc <[xfs]xfs_alloc_lookup+12c/3c0>
Trace; f899b114 <[xfs]xfs_free_ag_extent+254/770>
Trace; f899c470 <[xfs]xfs_free_extent+f0/100>
Trace; f8a09ff8 <[xfs]kmem_zone_zalloc+108/130>
Trace; f89f0d58 <[xfs]xfs_trans_get_efd+38/50>
Trace; f89ae8c8 <[xfs]xfs_bmap_finish+138/1d0>
Trace; f89d7da3 <[xfs]xfs_itruncate_finish+213/440>
Trace; f89f6623 <[xfs]xfs_inactive_free_eofblocks+293/2e0>
Trace; f89f6d4e <[xfs]xfs_release+9e/f0>
Trace; f89fc895 <[xfs]xfs_refcache_purge_some+e5/110>
Trace; f89f3c88 <[xfs]xfs_syncsub+258/330>
Trace; f89f2fc9 <[xfs]xfs_sync+29/30>
Trace; f8a05883 <[xfs]vfs_sync+43/50>
Trace; f8a0496a <[xfs]syncd+ca/f0>
Trace; c010752e <arch_kernel_thread+2e/40>
Trace; f8a048a0 <[xfs]syncd+0/f0>

2 warnings issued.  Results may not be reliable.

> >
> > I had to reboot system and run xfs_repair. it looks like xfs_repair does
> > not fix all problems, I have tried to run it several times and every time
> > had the same output.
> this may be normal, the first thing repair does is unlink lost+found, and
> then rediscovers everything in it.  rename lost+found and try again,
> things should be fine.  if you still see unrepaired errors please
> post the complete output of repair (or send directly to me if it's
> overly long)

I had to recreate file system and put it back to production. So, I can't try it.

> > I attached output from xfs_repair command ( I had to edit file to make it
> > readable ).
> hm, what was unreadable?

I mean, 313Mb of data is not quit easy to read (  > 6 millions line )

> -Eric

Marat Mukhitov         

<Prev in Thread] Current Thread [Next in Thread>