On 02/21/2011 12:45 PM, Steve Wilson wrote: > On 02/21/2011 09:54 AM, paul simpson wrote: >> hi fabricio, >> >> many thanks for your input. indeed i am using xfs - but that seems to be >> mentioned in the gluster docs without any mention of problems. we >> benchmarked xfs vs ext4 - and found that xfs to be much better at dealing >> with the bulk of our data - hi-def frames ~3-10M each - and large >> geometry/particle/volume files. 10M-200M. so, i'm keen to hear from >> anyone >> abotu xfs's suitability for gluster storage... >> >> as for file size; my understanding is that a distributed file system >> performance only really kicks in when your dealing with large>1M files. >> however, is dealing with small files meant to be unreliable with >> locking/access errors? >> > > We had trouble with reliability for small, actively-accessed files on a > distribute-replicate volume in both GlusterFS 3.11 and 3.12. It seems > that the replicated servers would eventually get out of sync with each > other on these kinds of files. For a while, we dropped replication and > only ran the volume as distributed. This has worked reliably for the > past week or so without any errors that we were seeing before: no such > file, invalid argument, etc. Steve: As a sanity check, do test your date stamps across the servers. We found *significant* issues when they drifted. -- Joseph Landman, Ph.D Founder and CEO Scalable Informatics Inc. email: landman at scalableinformatics.com web : http://scalableinformatics.com http://scalableinformatics.com/sicluster phone: +1 734 786 8423 x121 fax : +1 866 888 3112 cell : +1 734 612 4615
