Re: [Lustre-discuss] zero length files on lustre

2014-10-08 Thread anthony . david 


On 10/09/2014 11:50 AM, Anjana Kar wrote:
After some low level SAS expander firmware updates on our JetStor 
shelves,
the lustre filesystem is reporting several zero length files. The 
raids show

normal status, and all OSTs mount on the OSS nodes w/o problems. The
clients also mount the filesystem. Read/write tests from the clients 
report

no errors, but several files which had content before are showing up zero
length. These files appear to be on different OSTs so it's not on any 
particular

shelf. Rebooting the servers and a few clients hasn't shown any change.
Servers are running CentOS 5.5, lustre 1.8.3.


Hi Anjana

Have you looked at this bug? It might be related.

https://projectlava.xyratex.com/show_bug.cgi?id=24458

Cheers
Anthony


Could low level disk settings cause lustre to report zero length files?
There have been no software changes on these systems, so not sure
what could be causing this. Any thoughts?

Thanks,
-Anjana Kar
 Pittsburgh Supercomputing Center
 k...@psc.edu
___
Lustre-discuss mailing list
Lustre-discuss@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss


___
Lustre-discuss mailing list
Lustre-discuss@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss

[Lustre-discuss] zero length files on lustre

2014-10-08 Thread Anjana Kar 

After some low level SAS expander firmware updates on our JetStor shelves,
the lustre filesystem is reporting several zero length files. The raids show
normal status, and all OSTs mount on the OSS nodes w/o problems. The
clients also mount the filesystem. Read/write tests from the clients report
no errors, but several files which had content before are showing up zero
length. These files appear to be on different OSTs so it's not on any 
particular

shelf. Rebooting the servers and a few clients hasn't shown any change.
Servers are running CentOS 5.5, lustre 1.8.3.

Could low level disk settings cause lustre to report zero length files?
There have been no software changes on these systems, so not sure
what could be causing this. Any thoughts?

Thanks,
-Anjana Kar
 Pittsburgh Supercomputing Center
 k...@psc.edu
___
Lustre-discuss mailing list
Lustre-discuss@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss

Re: [Lustre-discuss] RPC packing

2014-10-08 Thread Dilger, Andreas 
It is also worthwhile to note that Lustre does not actually pack the 1MB or 4MB 
of data into the RPC request, but only a scatter-gather list of the data 
buffers and offsets. The data itself is transferred via RDMA.

Cheers, Andreas

On Oct 7, 2014, at 9:05, "teng wang" 
mailto:tzw0...@gmail.com>> wrote:

Hi,

 I know the RPC will pack the
client's data into 4MB buffer and transfer. The question is does the
client pack the writes even the writes on the client belong
to noncontiguous extents of the same file? Say, 1 client is issuing
two writes on two noncontiguous extents of the same file, the first write
from offset 0 to 16KB-1, the second write from 32KB to
48KB-1,  apparently these two writes are noncontiguous, so
will the two writes still be packed into the 4MB buffer?

Charles
___
Lustre-discuss mailing list
Lustre-discuss@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss
___
Lustre-discuss mailing list
Lustre-discuss@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss