I'm trying to run some tests using different aggregate sizes for write-behind. After mounting and unmount the glusterfs filesystem several times (with a change to the configuration), I start getting "Transport endpoint is not connected". It looks like shutting down glusterfs and removing the fuse module fixes the problem. Any ideas? Info: glusterfs-1.3.9-1 glusterfs-devel-1.3.9-1 fuse-2.7.4-1.el5.rf kernel-2.6.18-92.1.18.el5.centos.plus -- Matt
Matt, There might be a configuration error which is not letting the filesystem mount. Please check the log files if it has more details. Thanks, avati 2008/12/15 Matt Lawrence <matt.lawrence at tamu.edu>:> I'm trying to run some tests using different aggregate sizes for > write-behind. After mounting and unmount the glusterfs filesystem > several times (with a change to the configuration), I start getting > "Transport endpoint is not connected". It looks like shutting down > glusterfs and removing the fuse module fixes the problem. Any ideas? > > Info: > glusterfs-1.3.9-1 > glusterfs-devel-1.3.9-1 > fuse-2.7.4-1.el5.rf > kernel-2.6.18-92.1.18.el5.centos.plus > > -- Matt > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users >
Anand Avati wrote:> Matt, > There might be a configuration error which is not letting the > filesystem mount. Please check the log files if it has more details. > >The server log has a bunch of lines like: 2008-12-14 20:46:52 E [server-protocol.c:186:generic_reply] server: transport_writev failed 2008-12-14 20:46:52 E [server-protocol.c:186:generic_reply] server: transport_writev failed While it may be possible there is a configuration error, since the client configuration file is generated in the test script with only the size of the aggregate changing, I sort of doubt it. It does work some of the time. Any other ideas? -- Matt
Matt, With the available info we are not able to figure out what might be causing the problem. I just tried your vol files and it worked fine. Does glusterfs work when you start it and later start failing? Any idea on what triggers the bug? Is it possible for you to check with the latest glusterfs codebase (1.4)? Thanks Krishna On Tue, Dec 16, 2008 at 1:34 AM, Matt Lawrence <matt.lawrence at tamu.edu> wrote:> Looking in /var/log/messages, I see: > > Dec 15 13:59:14 radarnfs kernel: fuse init (API version 7.7) > Dec 15 14:00:36 radarnfs kernel: glusterfs[3365] trap divide error > rip:2ae25c98bfbc rsp:7fff4e33d910 error:0 > > > Probably not a good thing. > > -- Matt > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users >