gluster no longer serving all of its files

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

 



Pat,
     I looked at one more log and the same pattern follows:

pranithk @ ~/Downloads/pat-mit/server
15:44:04 :) $ grep -w disconnected gdata.log | awk '{print $5, $6}' | 
sort | uniq -c
       1 0-gdata-client-0: disconnected
       2 0-gdata-client-1: disconnected
     608 0-gdata-client-2: disconnected <--------
       4 gdata-client-0: disconnected
       4 gdata-client-1: disconnected
      22 gdata-client-2: disconnected

Pranith

On 02/01/2012 03:21 PM, Pranith Kumar K wrote:
> hi Pat,
>     It seems that brick is disconnecting frequently:
> pranithk @ ~/Downloads/pat-mit/old
> 14:48:39 :) $ grep -w disconnected gdata.log | awk '{print $5, $6}' | 
> sort | uniq -c
>       5 0-gdata-client-0: disconnected
>       6 0-gdata-client-1: disconnected
>     308 0-gdata-client-2: disconnected <<----- It disconnected 308 
> times where as the others < 10
>       2 gdata-client-0: disconnected
>       1 gdata-client-1: disconnected
>       8 gdata-client-2: disconnected
>
> You were mentioning that you tried doing self-heal, but I dont see any 
> replicate configuration in your setup. It is a plain distribute setup. 
> Am I missing something?.
>
> This is the file I looked into 
> http://mseas.mit.edu/download/phaley/GlusterUsers/Old/client_logs/gdata.log
>
> Pranith.
>
> On 01/30/2012 11:47 PM, Pat Haley wrote:
>>
>> Hi,
>>
>> We recently upgraded our version of gluster from 3.1.4 to
>> 3.2.5. Shortly after I noticed that gluster was not serving
>> up all of its files, that is if I log onto one of the individual
>> bricks and do an ls of the underlying nfs directories, I can
>> see files that I do not see if do an ls from a client node
>> in the equivalent gluster directory.
>>
>> We may have caused this by also running a script that went
>> through and removed pointer files (this had been how we had
>> been dealing with bad pointers under gluster 3.1.4).
>>
>> So far, we have tried
>>    - running a gluster self-heal under version 3.2.5
>>    - rolling back to version 3.1.4
>>    - running a gluster self-heal under version 3.1.4
>>    - running a rebalance under version 3.1.4
>> None of these have solved the problem.  The one other
>> piece of data we have is that all the files which are
>> not being served appear to reside on a single brick
>> (at least every file we missed so far has been on
>>  that server and we have not found one missing from
>>  the other servers).
>>
>> Any advice you can give us would be greatly appreciated.
>>
>> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
>> Pat Haley                          Email:  phaley at mit.edu
>> Center for Ocean Engineering       Phone:  (617) 253-6824
>> Dept. of Mechanical Engineering    Fax:    (617) 253-8125
>> MIT, Room 5-222B                   http://web.mit.edu/phaley/www/
>> 77 Massachusetts Avenue
>> Cambridge, MA  02139-4301
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users



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

  Powered by Linux