[Gluster-users] Client and server file "view", different results?! Client can't see the right file.

[Martin Schenker]

No, we had these issues before already, running on 3.1.3 
Only the system load has gone up a lot in the meantime...

Best, Martin

-----Original Message-----
From: Mohit Anchlia [mailto:mohitanchlia at gmail.com] 
Sent: Wednesday, May 18, 2011 10:36 PM
To: Martin Schenker
Cc: gluster-users at gluster.org
Subject: Re: [Gluster-users] Client and server file "view", different
results?! Client can't see the right file.

So you started seeing this issue after rolling it back to 3.1.3?

On Wed, May 18, 2011 at 1:30 PM, Martin Schenker
<martin.schenker at profitbricks.com> wrote:
> We're running 3.1.3. we had a brief test of 3.2.0 and rolled back to 3.1.3
> by reinstalling the Debian package.
>
> 0 root at pserver12:~ # gluster volume info all
>
> Volume Name: storage0
> Type: Distributed-Replicate
> Status: Started
> Number of Bricks: 4 x 2 = 8
> Transport-type: rdma
> Bricks:
> Brick1: de-dc1-c1-pserver3:/mnt/gluster/brick0/storage
> Brick2: de-dc1-c1-pserver5:/mnt/gluster/brick0/storage
> Brick3: de-dc1-c1-pserver3:/mnt/gluster/brick1/storage
> Brick4: de-dc1-c1-pserver5:/mnt/gluster/brick1/storage
> Brick5: de-dc1-c1-pserver12:/mnt/gluster/brick0/storage
> Brick6: de-dc1-c1-pserver13:/mnt/gluster/brick0/storage
> Brick7: de-dc1-c1-pserver12:/mnt/gluster/brick1/storage
> Brick8: de-dc1-c1-pserver13:/mnt/gluster/brick1/storage
> Options Reconfigured:
> network.ping-timeout: 5
> nfs.disable: on
> performance.cache-size: 4096MB
>
> Best, Martin
>
> -----Original Message-----
> From: Mohit Anchlia [mailto:mohitanchlia at gmail.com]
> Sent: Wednesday, May 18, 2011 9:43 PM
> To: Martin Schenker
> Cc: gluster-users at gluster.org
> Subject: Re: [Gluster-users] Client and server file "view", different
> results?! Client can't see the right file.
>
> Which version are you running? Can you also post output from volume info?
>
> Meanwhile, anyone from dev want to answer??
>
> On Wed, May 18, 2011 at 1:53 AM, Martin Schenker
> <martin.schenker at profitbricks.com> wrote:
>> Here is another occurrence:
>>
>> The file 20819 is shown twice, different timestamps and attributes. 0
>> filesize on pserver3, outdated on pserver5, just 12&13 seems to be in
> sync.
>> So what's going on?
>>
>>
>> 0 root at de-dc1-c1-pserver13:~ # ls -al
>>
>
/opt/profitbricks/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/h
>> dd-images/2081*
>> -rwxrwx--- 1 libvirt-qemu kvm 53687091200 May 18 08:44
>>
>
/opt/profitbricks/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/h
>> dd-images/20819
>> -rwxrwx--- 1 libvirt-qemu kvm 53687091200 May 18 08:44
>>
>
/opt/profitbricks/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/h
>> dd-images/20819
>>
>> 0 root at de-dc1-c1-pserver3:~ # find /mnt/gluster/brick?/ -name 20819 |
> xargs
>> -i ls -al {}
>> -rwxrwx--- 1 libvirt-qemu vcb 0 May 14 17:00
>>
>
/mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> 0 root at de-dc1-c1-pserver3:~ # getfattr -dm -
>>
>
/mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> getfattr: Removing leading '/' from absolute path names
>> # file:
>>
>
mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/
>> hdd-images/20819
>> trusted.gfid=0sa5/rvjUUQ3ibSf32O3izOw==
>>
>> 0 root at pserver5:~ # find /mnt/gluster/brick?/ -name 20819 | xargs -i ls
> -al
>> {}
>> -rwxrwx--- 1 libvirt-qemu vcb 53687091200 May 14 17:00
>>
>
/mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> 0 root at pserver5:~ # getfattr -dm -
>>
>
/mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> getfattr: Removing leading '/' from absolute path names
>> # file:
>>
>
mnt/gluster/brick0/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/
>> hdd-images/20819
>> trusted.afr.storage0-client-0=0sAAAAAgAAAAIAAAAA
>> trusted.afr.storage0-client-1=0sAAAAAAAAAAAAAAAA
>> trusted.gfid=0sa5/rvjUUQ3ibSf32O3izOw==
>>
>> 0 root at pserver12:~ # find /mnt/gluster/brick?/ -name 20819 | xargs -i ls
> -al
>> {}
>> -rwxrwx--- 1 libvirt-qemu kvm 53687091200 May 18 08:41
>>
>
/mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> 0 root at pserver12:~ # getfattr -dm -
>>
>
/mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> getfattr: Removing leading '/' from absolute path names
>> # file:
>>
>
mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/
>> hdd-images/20819
>> trusted.afr.storage0-client-6=0sAAAAAAAAAAAAAAAA
>> trusted.afr.storage0-client-7=0sAAAAAAAAAAAAAAAA
>> trusted.gfid=0sa5/rvjUUQ3ibSf32O3izOw==
>> trusted.glusterfs.dht.linkto="storage0-replicate-0
>>
>> 0 root at de-dc1-c1-pserver13:~ # find /mnt/gluster/brick?/ -name 20819 |
> xargs
>> -i ls -al {}
>> -rwxrwx--- 1 libvirt-qemu kvm 53687091200 May 18 08:39
>>
>
/mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> 0 root at de-dc1-c1-pserver13:~ # getfattr -dm -
>>
>
/mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef
>> /hdd-images/20819
>> getfattr: Removing leading '/' from absolute path names
>> # file:
>>
>
mnt/gluster/brick1/storage/images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/
>> hdd-images/20819
>> trusted.afr.storage0-client-6=0sAAAAAAAAAAAAAAAA
>> trusted.afr.storage0-client-7=0sAAAAAAAAAAAAAAAA
>> trusted.gfid=0sa5/rvjUUQ3ibSf32O3izOw==
>> trusted.glusterfs.dht.linkto="storage0-replicate-0
>>
>> Only entrance in log file on pserver5, no references in the other three
>> logs/servers:
>>
>> 0 root at pserver5:~ # grep 20819
>> /var/log/glusterfs/opt-profitbricks-storage.log
>> [2011-05-17 20:37:30.52535] I
> [client-handshake.c:407:client3_1_reopen_cbk]
>> 0-storage0-client-7: reopen on
>> /images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/hdd-images/20819
> succeeded
>> (remote-fd = 6)
>> [2011-05-17 20:37:34.824934] I [afr-open.c:435:afr_openfd_sh]
>> 0-storage0-replicate-3:  data self-heal triggered. path:
>> /images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/hdd-images/20819,
> reason:
>> Replicate up down flush, data lock is held
>> [2011-05-17 20:37:34.825557] E
>> [afr-self-heal-common.c:1214:sh_missing_entries_create]
>> 0-storage0-replicate-3: no missing files -
>> /images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/hdd-images/20819.
>> proceeding to metadata check
>> [2011-05-17 21:08:59.241203] I
>> [afr-self-heal-algorithm.c:526:sh_diff_loop_driver_done]
>> 0-storage0-replicate-3: diff self-heal on
>> /images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/hdd-images/20819: 6
> blocks
>> of 409600 were different (0.00%)
>> [2011-05-17 21:08:59.275873] I
>> [afr-self-heal-common.c:1527:afr_self_heal_completion_cbk]
>> 0-storage0-replicate-3: background  data self-heal completed on
>> /images/2078/ebb83b05-3a83-9d18-ad8f-8542864da6ef/hdd-images/20819
>>
>>
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>>
>
>