On 25/11/12 16:16, Josselin Mouette wrote: > Le dimanche 21 octobre 2012 à 21:34 +0200, Daniel Pocock a écrit : > >> I am raising the severity to `serious' because this hammers the NFS >> fileserver and causes severe performance degradation for users logged in >> with NFS home directories. >> >> Similar issue reported in Fedora, unfixed there too apparently: >> >> https://bugzilla.redhat.com/show_bug.cgi?id=561904 >> >> The explanation at the bottom suggests that it is caused by multiple >> processes accessing the redo log concurrently, particularly if there are >> multiple NFS clients logged in as the same user presumably. >> > This looks like a very serious design issue with gvfs indeed. > > Has this been forwarded to bugzilla.gnome.org or discussed on upstream > mailing lists somewhere? >
I have only reported it in Debian BTS. I haven't had time to investigate the issue more thoroughly, although I did verify that it appears the same (e.g. monitoring NFS client stats, killing other processes, and then killing this process to verify the NFS behavior returns to a normal level) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
