[Gluster-users] "fd not open on any subvolume" causes root ownership

Dan Bretherton d.a.bretherton at reading.ac.uk
Sat Aug 6 12:39:39 UTC 2011


Hello Pranith,
I believe the user generated the files using the Climate Data Operators 
(CDO) tools (https://code.zmaw.de/projects/cdo).  If it would help, I 
will ask the user if there is an easy way to reproduce the problem, 
ideally without the need for a large and complicated input data set.
Regards
Dan.

-- 
Mr. D.A. Bretherton
Computer System Manager
Environmental Systems Science Centre
Harry Pitt Building
3 Earley Gate
University of Reading
Reading, RG6 6AL
UK

Tel. +44 118 378 5205
Fax: +44 118 378 6413


On 05/08/11 14:17, Pranith Kumar K wrote:
> hi Dan,
>      We tried reproducing it in our local setup but the user 
> ownership/permissions are not changing when this situation is hit, 
> there must be some other reason why this would have happened. Could 
> you let us know what all operations were going on before it went into 
> this situation.
>
>     Ignore this log, it is a misleading log that will be removed in 
> the future releases.
>
> Pranith
>
> On 08/04/2011 05:46 PM, Dan Bretherton wrote:
>> Hello All-
>> I have had a spate of these errors in one of my volumes recently, 
>> with GlusterFS version 3.2.2.
>>
>> [2011-08-02 16:14:38.211400] W [afr-open.c:624:afr_openfd_flush] 
>> 0-atmos-replicate-7: fd not open on any subvolume 0x2aaaaf51e174 
>> (/users/rle/CMIP5/NCC/NorESM1-M/rcp45/r1i1p1/ua_6hrPlev_NorESM1-M_rcp45_r1i1p1_2182010100-2182123118.nc) 
>>
>>
>> The files have root ownership (which is not correct) but are 
>> otherwise intact.  The errors occur when the files are being written 
>> and the problem is not restricted to one client or one server.  I 
>> have seen this error in GlusterFS client logs and in nfs.log.  I am 
>> not running any rebalance operations at the moment.
>>
>> Does anybody know what causes this or how to stop it?  I hope I don't 
>> have to go back to doing "chown -R" in people's directories regularly 
>> to correct ownership problems.
>>
>> Regards
>> Dan.
>




More information about the Gluster-users mailing list