Re: Unable to re-export aufs with nfsv3
Great! You hit the mark. Just an update, I workaround exporting also with nfsv3 protocol inserting a layer over the original nfs branch. At moment for the overlay I used fuse.bindfs. I done some test and I expecienced only a crash on process nfsd with heavy load, performances are non good but suffients. Bye In data mercoledì 11 marzo 2020 15:35:48 CET, J. R. Okajima ha scritto: > Leonardo Secci: > > It seems that this configuration works with NFSv4, performance are not > > grea= t=20 > > but this is not the focus. > > I've found a commit from v4.13, > > commit 20fa19027286983ab2734b5910c4a687436e0c31 > Author: Peng Tao > Date: Thu Jun 29 06:34:53 2017 -0700 > > nfs: add export operations > > This support for opening files on NFS by file handle, both through the > open_by_handle syscall, and for re-exporting NFS (for example using a > different version). The support is very basic for now, as each open by > handle will have to do an NFSv4 open operation on the wire. In the > future this will hopefully be mitigated by an open file cache, as well > as various optimizations in NFS for this specific case. > > Waoh, I didn't know that. But it explains your case well. > NFS client can export nfs-mount on himself, but this re-exporting is > supported by nfs4 only. > > > J. R. Okajima -- -- Leonardo Secci mailto:leonardo.se...@unirel.com UniRel s.r.l. Via Arno località Osmannoro Sesto Fiorentino - 50019 Firenze Web: http://www.unirel.com -- Avviso di riservatezza Le informazioni contenute nella presente comunicazione e i relativi allegati possono essere riservate e sono, comunque, destinate esclusivamente alle persone o alla Società sopraindicati. La diffusione, distribuzione e/o copiatura del documento trasmesso da parte di qualsiasi soggetto diverso dal destinatario è proibita, sia ai sensi dell’art. 616 c.p., che ai sensi del D.Lgs. 196/2003. Se avete ricevuto questo messaggio per errore, Vi preghiamo di distruggerlo o di informarci immediatamente. The information in this e-mail (which includes any files transmitted with it)is confidential and may also be legally privileged. It is intended for the addressee only. Access to this e-mail by anyone else is unauthorised. It is not to be relied upon by any person other than the addresse, except with our prior written approval. If no such approval is given, we will not accept any liability (in negligense or otherwise) rising from any third acting, or refraining from acting on such information. Unauthorised recipients are required to mantain confidentiality. If you have received this e-mail in error please notify immediately, destroy any copies ad delete it from your computer system. Any use, dissemination, forwarding, printing or copying of this e-mail is prohibited. Copyright in this e-mail and any document created by us will be remain vested in us and will not be transferred to you. We assert the right to be identified as the author of and to object to any misuses of the contents of this e-mail or such documents.
Re: Unable to re-export aufs with nfsv3
Leonardo Secci: > It seems that this configuration works with NFSv4, performance are not grea= > t=20 > but this is not the focus. I've found a commit from v4.13, commit 20fa19027286983ab2734b5910c4a687436e0c31 Author: Peng Tao Date: Thu Jun 29 06:34:53 2017 -0700 nfs: add export operations This support for opening files on NFS by file handle, both through the open_by_handle syscall, and for re-exporting NFS (for example using a different version). The support is very basic for now, as each open by handle will have to do an NFSv4 open operation on the wire. In the future this will hopefully be mitigated by an open file cache, as well as various optimizations in NFS for this specific case. Waoh, I didn't know that. But it explains your case well. NFS client can export nfs-mount on himself, but this re-exporting is supported by nfs4 only. J. R. Okajima
Re: Unable to re-export aufs with nfsv3
Hi and thank you for support, I tried to run pts/disk of Phoronix Test Suite in the mounted nfsv4 partition and I not experienced any failure, after that I executed also iozone without errors. But this test targets only on writable branch, so I try to stress also the read only nfs branch with some concurrent readers, same results. It seems that this configuration works with NFSv4, performance are not great but this is not the focus. Bye In data mercoledì 11 marzo 2020 09:42:58 CET, J. R. Okajima ha scritto: > Hello Leonardo, > > Leonardo Secci: > > I have an issue exporting an aufs filesystem usiing nfsv3, I can mount NFS > > export with success but I get an error reading a file or a directory. > > > > Into syslog I see: > > [ 183.774819] aufs aufs_encode_fh:763:nfsd[1467]: nfs encode_fh failed > > > > > > [CLIENTc7]# ls /mail/01/BOX1/Maildir/ > > > > [CLIENTc7]# cat /mail/01/BOX1/Maildir/maildirsize cat: > > /mail/01/BOX1/Maildir/maildirsize: Input/output error > > According to your additional info, I can see that > - you have two aufs mounts on your nfs server SERVERc7, /mail/overlay/0[13] > + /mail/overlay/01 = /mail/new/01 + /mail/prod/01 > + /mail/overlay/03 = /mail/new/03 + /mail/prod/03 > - /mail/new/0[13] are xfs > - /mail/prod/0[13] are nfs > > The aufs manual describes some limitations about "Exporting Aufs via > NFS", > > (quote) > > .SH Exporting Aufs via NFS > Aufs is supporting NFS\-exporting. > Since aufs has no actual block device, you need to add NFS `fsid' option at > exporting. Refer to the manual of NFS about the detail of this option. > > There are some limitations or requirements. > .RS > .Bu > The branch filesystem must support NFS\-exporting. > > > Your branch fs /mail/prod/0[13] are nfs and they don't support > NFS\-exporting itself. > > I don't know why /mail/03, nfs4 on CLIENTc7. > Is it stably working? Is it OK if you try giving some heavy workloads > on CLIENTc7? > > > J. R. Okajima
Re: Unable to re-export aufs with nfsv3
Hello Leonardo, Leonardo Secci: > I have an issue exporting an aufs filesystem usiing nfsv3, I can mount NFS > export with > success but I get an error reading a file or a directory. > > Into syslog I see: > [ 183.774819] aufs aufs_encode_fh:763:nfsd[1467]: nfs encode_fh failed > > > [CLIENTc7]# ls /mail/01/BOX1/Maildir/ > > [CLIENTc7]# cat /mail/01/BOX1/Maildir/maildirsize cat: > /mail/01/BOX1/Maildir/maildirsize: > Input/output error ::: According to your additional info, I can see that - you have two aufs mounts on your nfs server SERVERc7, /mail/overlay/0[13] + /mail/overlay/01 = /mail/new/01 + /mail/prod/01 + /mail/overlay/03 = /mail/new/03 + /mail/prod/03 - /mail/new/0[13] are xfs - /mail/prod/0[13] are nfs The aufs manual describes some limitations about "Exporting Aufs via NFS", (quote) .SH Exporting Aufs via NFS Aufs is supporting NFS\-exporting. Since aufs has no actual block device, you need to add NFS `fsid' option at exporting. Refer to the manual of NFS about the detail of this option. There are some limitations or requirements. .RS .Bu The branch filesystem must support NFS\-exporting. Your branch fs /mail/prod/0[13] are nfs and they don't support NFS\-exporting itself. I don't know why /mail/03, nfs4 on CLIENTc7. Is it stably working? Is it OK if you try giving some heavy workloads on CLIENTc7? J. R. Okajima
