On 2013/12/17 1:52 PM, "Kumar, Amit" <ahkumar(a)mail.smu.edu> wrote:
Finally found an answer. Scripts were writing to /tmp as an
intermediate
space that was limited and hence the error. Sorry for the trouble.
Thanks for the follow-up. It is always good to know what solved a
problem in the end (even if not Lustre related), especially for later
users searching the list archives.
Cheers, Andreas
-----Original Message-----
From: HPDD-discuss [mailto:hpdd-discuss-bounces@lists.01.org] On Behalf
Of Kumar, Amit
Sent: Monday, December 16, 2013 12:01 PM
To: Mohr Jr, Richard Frank (Rick Mohr)
Cc: hpdd-discuss(a)lists.01.org
Subject: Re: [HPDD-discuss] False: No Space Left on device or inode
limitation reached?
Hi Rick,
No quotas.
But I will try the inode and getstripe suggestion at error, and hopefully
find something.
Thank you for the suggestion.
Amit
-----Original Message-----
From: Mohr Jr, Richard Frank (Rick Mohr) [mailto:rmohr@utk.edu]
Sent: Monday, December 16, 2013 11:58 AM
To: Kumar, Amit
Cc: hpdd-discuss(a)lists.01.org
Subject: Re: [HPDD-discuss] False: No Space Left on device or inode
limitation reached?
On Dec 16, 2013, at 12:04 PM, "Kumar, Amit" <ahkumar(a)mail.smu.edu>
wrote:
> Thank you for your response. Yes I checked "lfs df" and all OST's are
>less than 77% .
You can also run "lfs df -i" to check the inode usage on each ost/mdt and
see the maximum number of inodes available.
> Interestingly during multiple runs this error is seen at different
>point of time.
> First time error resulted in after generating 3.5 million files, second
>time after generating 2.5 million files, hence not able to really tell
>what exactly could be causing this.
Do you have a way to tell which file was being written when the error is
encountered? If so, you could always run "lfs getstripe" to check the
striping information and see if there is anything in common (like a
specific ost always being used).
> Any other thoughts.
Do you use quotas on the file system?
--
Rick Mohr
Senior HPC System Administrator
National Institute for Computational Sciences
http://www.nics.tennessee.edu
_______________________________________________
HPDD-discuss mailing list
HPDD-discuss(a)lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss
_______________________________________________
HPDD-discuss mailing list
HPDD-discuss(a)lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss
Cheers, Andreas
--
Andreas Dilger
Lustre Software Architect
Intel High Performance Data Division