…
> The time variation is caused cache coherence. when client has valid information
> in its cache, 'stat' operation will be fast. Otherwise the client need to send
> request to MDS and wait for reply, which will be slow.
> The time variation is caused cache coherence. when client has valid information
> in its cache, 'stat' operation will be fast. Otherwise the client need to send
> request to MDS and wait for reply, which will be slow.
This sounds like the behavior I had with CephFS giving me question marks. When I had a directory with a large amount of files in it and the first ls -la took a while to populate and ended with some unknown stats. The second time I did an ls -la it ran quick with no question marks. My inquiry was if there is a timeout that could occur? since it has to go ask the mds on a different machine it seems plausible that the full response is not coming back in time or fails to get all stats at some point.
I could test this more; is there a command or proccess I can perform to flush the ceph-fuse cache?
Thanks,
Scott
On Fri, Mar 13, 2015 at 1:49 PM Francois Lafont <flafdivers@xxxxxxx> wrote:
Hi,
Yan, Zheng wrote :
>> http://tracker.ceph.com/issues/11059
>>
>
> It's a bug in ACL code, I have updated http://tracker.ceph.com/issues/11059
Ok, thanks. I have seen and I will answer quickly. ;)
>> I'm still surprised by such times. For instance, It seems to me
>> that, with a mounted nfs share, commands like "ls -la" are very
>> fast in comparison (with a directory which contains the same number
>> of files). Can anyone explain to me why there is a such difference
>> between the nfs case and the cephfs case? This is absolutely not a
>> criticism but it's just to understand the concepts that come into
>> play. In the case of "ls -al" ie just reading (it is assumed that
>> there is no writing on the directory), the nfs and the cephfs cases
>> seem to me very similar: the client just requests a stat on each file
>> in the directory. Am I wrong?
>
> NFS has no cache coherence mechanism. It can't guarantee one client always
> see other client's change.
Ah ok, I didn't know that. Indeed, now I understand that can generate
performance impact.
> The time variation is caused cache coherence. when client has valid information
> in its cache, 'stat' operation will be fast. Otherwise the client need to send
> request to MDS and wait for reply, which will be slow.
Ok, thanks a lot for your explanations.
Regards.
--
François Lafont
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
