Re: GlusterFS core dump (v1.2.3)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Ian,
this most likely seems to be an installation issue. you seem to have two
installations in /usr/local and /usr. let me list the files you need to
delete to be sure

${prefix}/sbin/glusterfs
${prefix}/sbin/glusterfsd
${prefix}/lib/libglusterfs.so.*
${prefix}/lib/glusterfs -rf

make sure you clean out stuff from both /usr and /usr/local prefixes and do
a fresh install. also ensure that the export directories are having
consistant directory structures on all export macihnes (the newer versions
will automatically fix this error)

thanks,
avati

2007/6/25, Ian Grimstead <I.J.Grimstead@xxxxxxxxxxxxxxxx>:

Hi Avati,

Thanks for the v.prompt response; I've not tried on an alt. / clean
system,
I've only got sane access to the cluster I'm trying it on...

To uninstall, I simply used "make uninstall" in the 1.2.3 build dir, but
looking in /usr/lib/glusterfs:

1.3.0-pre3    scheduler    transport    xlator

scheduler, transport & xlatar are all empty dirs, or contain empty dirs.
I've removed them from each machine for tidyness.

Also had old libs loitering in /usr/local/lib/glusterfs - removed these...
got suspicious and trawled my entire HDs for gluster. Steam-cleaned
anything
left after "make uninstall".

gluster--mainline--2.4 checked to new work area... "make install" ran on
each machine (gluster src mounted over NFS).

Spec file - do you mean the .vol? I've attached them both (client &
server).
Machines are:

n1, n3, n5: glusterfs daemon, local directory hosted
n7: localdir and ext. usb disc hosted (alt. vol file)

n2, n4, n6, n8: no drives hosted; these will mirror n1,3,5,7 using the
replication setup once the basics are stable.

Then n1-8 are all clients, along with "head1". Everything's running RHEL
(uh) 4, example uname from n1:

Linux n1 2.6.9-42.ELsmp #1 SMP Wed Jul 12 23:32:02 EDT 2006 x86_64 x86_64
x86_64 GNU/Linux

n1-8 all dual CPU, dual core => 4 opteron cores.
head1 is a Sun x4600, so 4 CPUs, all dual core (=> 8 cores)

Compiled glusterfs *without* a "--disable-fuse-client" or
"--disable-server", as 95% of machines will be both servers and clients,
so
a single compilation is then installed (via NFS share) to all machines.

Servers launched with:
LD_LIBRARY_PATH=/usr/local/lib:/usr/lib glusterfsd -f
/etc/glusterfs/glusterfs-server.vol (path req'd to pick up fuse etc)

USB hosting servers launched with:
LD_LIBRARY_PATH=/usr/local/lib:/usr/lib glusterfsd -f
/etc/glusterfs/glusterfs-server-usb.vol

Clients launched with:

modprobe fuse
LD_LIBRARY_PATH=/usr/local/lib:/usr/lib glusterfs -f
/etc/glusterfs/glusterfs-client.vol /media/glusterfs

I think that's all the spec... :o)

Now, with this install... Glusterfs has packed up on me! For instance, I
mount glusterfs on /mount/glusterfs.

ls -l  /media/glusterfs
total 24
drwxr-xr-x  3 root root 4096 May 30 14:57 .
drwxr-xr-x  6 root root 4096 Jun 25 14:26 ..
drwxr-xr-x  4 root root 4096 Jun 11 13:46 home

Looks ok, but:
ls -l /media/glusterfs/home
total 0

Ah. If I "cd /media/glusterfs/home", I can then cd into userdirs (e.g.
"guest"), and they work fine. "ls" doesn't work though...

I've disabled writebehind, readahead and statprefetch to see if anything
was
interfering... No change. Status attached as per vol files (statprefetch
disabled); also attached log from client-debug...

Help! :o)

Ian



________________________________

        From: anand.avati@xxxxxxxxx [mailto:anand.avati@xxxxxxxxx] On
Behalf
Of Anand Avati
        Sent: 25 June 2007 13:01
        To: Ian Grimstead
        Cc: gluster-devel@xxxxxxxxxx
        Subject: Re: GlusterFS core dump (v1.2.3)


        Can you please attach the spec file? probably some stale binaries
have got left over from the previous build.. what were the steps you did
to
uninstall glusterfs? does it work on a differnt system which had no traces
of 1.2.3 ?

        thanks,
        avati


        2007/6/25, Ian Grimstead <I.J.Grimstead@xxxxxxxxxxxxxxxx>:

                I think something's up - I have just uninstalled 1.2.3 on
each node, then obtained & built glusterfs--mainline--2.4; looks like you
need to run autogen.sh to produce the "configure" script?

                Did that, installed... now I'm not getting any directories
under my glusterfs mount, apart from a single (and empty) "/home". So I
have:

                /media/glusterfs => gluster FS mount point

                ls -al /media/glusterfs gives:

                total 24
                drwxr-xr-x  3 root root 4096 May 30 14:57 .
                drwxr-xr-x  3 root root 4096 May 11 17:32 ..
                drwxr-xr-x  4 root root 4096 Jun 11 13:46 home

                But, ls -al /media/glusterfs/home gives:

                total 0

                Rather worrying. However, if I "cd" to a known directory,
it
works. This dir is also empty. If I "cat" a known file, it's there!

                Wierd. Log from client just shows "stat-prefetch:flush on:
/" 3 times after initial handshake success of servers...

                I've turned off prefetch caching on the client - no
difference. :o(

                Any idea what could have caused this?

                Ian



        --
        Anand V. Avati





--
Anand V. Avati


[Index of Archives]     [Gluster Users]     [Ceph Users]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux