On 12/05/2012 04:43 PM, Amon Ott wrote: > Am 04.12.2012 09:09, schrieb Yan, Zheng: >> These two patches fix MDS restart issues that may corrupt the fragstat accounting >> and cause undeletable empty directory. These two patches, together with the patch >> "mds: fix journaling issue regarding rstat accounting", have passed overnight >> "fsstress + restart MDS every 120s" test. No undeletable directory was found. >> >> These patches are also in: >> git://github.com/ukernel/ceph.git wip-mds > > Thank you very much for working on MDS issues, your work gave me some > hope that our very long standing crashes at MDS failover could be fixed > soon. Hi, Sorry for the delay > > Right now, we have our four node test cluster running with 0.55 and your > latest patches. Hopefully, today's stress test will work fine. > > One question: with your latest patches, is a multiple MDS setup stable > in your tests? So far, we only tried with max mds = 1. As our work load > writes many small files, some load sharing for MDS would be nice to > have. I would be willing to help with tests and reports, if needed. > No, multiple MDS setup is still unstable. Regards Yan, Zheng > Amon Ott > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
