On Mon, 11 Jul 2011, Fyodor Ustinov wrote: > On 07/11/2011 12:12 AM, Sage Weil wrote: > > Hmm, yeah the 'loaded dup inode' messages should never happen. They > > appear to be just files, though, not directories, so you can safely ignore > > them. The namespace repair will need to clean that up at some point. The > > recursive rstat errors you saw are probably just fallout from that, and > > can also be ignored. > I understand it correctly, that until not implemented the Feature # 86 I have > no way to fix this? Right. Well, this is going to come incrementally. We'll have a namespace scan that fixes some basic issues before there is a full-blown robust repair tool. > > I'm curious how you got into that state, though. Are you running a single > > or clustered mds? Is the workload purely amanda? > > Single mds. I have these clients: > 1. Amanda. Used cfuse. > 2. Gate. Used cfuse. NFS gate. > 3. One experimental server with many different workloads. > > WBR, > Fyodor. > > P.S. My friends have always said that I can find a bug in the software, even > where it can not be. Sometimes it interferes. :) > > I will try again to use ceph with BackupPC. Last time it was over "brain > explosion" in ceph. :) :) Okay. We're putting locking specific tests in place before trying to reproduce #1150 with Amanda, BTW. It sounds like we need an amanda-like workload though anyway, given the issues you're seeing. How old is this fs? How many times has it been upgraded? sage -- 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
