Re: process stuck in nfsfsync state

From: Joan Picanyol (lists-freebsd-stable_at_biaix.org)
Date: 10/25/04

  • Next message: Manuel Martini: "Dell CERC SATA RAID controller"
    Date: Mon, 25 Oct 2004 18:09:30 +0200
    To: Robert Watson <rwatson@freebsd.org>
    
    

    * Robert Watson <rwatson@freebsd.org> [20041025 14:24]:
    >
    > On Mon, 25 Oct 2004, Joan Picanyol wrote:
    >
    > > > Is there an response to the request? If not, that might suggest the
    > > > server is wedged, not the client. If you are willing to share the results
    > > > of a tcpdump -s 1500 -w <whatever> output from a few seconds during the
    > > > wedge, that would be very useful.
    > >
    > > Available at http://biaix.org/pk/debug/nfs/ These are from just after
    > > logging in to GNOME until gconfd-2 goes to nfsfsync, and the nfs server
    > > not responding messages start appearing.
    >

    [snip *much* appreciated detailed analysis]

    > So if possible, I might try some of the following:
    [...]
    > - I think someone already suggested disabling hardware checksumming, but
    > if you haven't tried that, it would be worth trying it.

    No difference.

    > - It would be useful to see if less complicated NFS meta-transactions than
    > "Start GTK" can trigger the problem. For example, doing a large dd to a
    > file in NFS, varying the blocksize to see if you can find useful
    > thresholds that trigger the problem. I see a lot of successful 512 byte
    > writes in the trace, but larger datagram sizes of 8192 for writes seem
    > to have problems.

    Now this is interesting:

    dd if=/dev/urandom of=/fs/bulk/mount/dummy bs=512 count=14

    wedges the NFS mount point 100% of the times. Lowering the count to 13
    doesn't reproduce the hang.

    An another possibly interesting data point is that NFS over TCP works
    ok. For this I'm particularly grateful, since I can now mount my /home
    fs and do my work.

    Am I the only one seeing this?

    tks

    -- 
    pica
    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"
    

  • Next message: Manuel Martini: "Dell CERC SATA RAID controller"

    Relevant Pages

    • [GIT] Please pull fixes for 2.6.23-rc5 nfs client
      ... commit 1b3b4a1a2deb7d3e5d66063bd76304d840c966b3 ... NFS: change NFS mount error return when hostname/pathname too long ... The hostname was getting truncated in the new text-based NFS mount API. ... Return the real error code so that callers of the new NFS ...
      (Linux-Kernel)
    • [HPADM] RE: -SUMMARY- NFS over TCP
      ... assuming you have cycled the NFS service down and back up after ... connections to/from nfsd in the tcp section, ... nfs mount: getaddr_nfs: oldserver: NFS service not responding(retry ...
      (HP-UX-Admin)
    • Re: NFS and mount_null
      ... > only one NFS cache and connection for all of them. ... > point, it takes a long time the first time, and then it's cached. ... you can't re-export an nfs mount. ...
      (comp.unix.bsd.freebsd.misc)
    • [SLE] 8 character hostname limit? Was Re: nfs mounting problem
      ... being an NFS specific problem and people not having NFS problems ... For two out of five 10.1 machines, ... On the SLES 9 NFS server, the filesystem is exported as follows (in ... NFS mount failures went away. ...
      (SuSE)
    • Re: Waiting for the Fuck-Up Fairy
      ... which won't boot because it needs to nfs mount something from bar ... NFS clusterfsck every time the power blipped. ... my boxes came up in the proper 30 seconds it ... took to boot, while the other groups got to wait the 1 hour for NFS ...
      (alt.sysadmin.recovery)