Hello! m.c.wilkins at massey.ac.nz pisze:>>> "transport endpoint not connected" after few minutes of usage ;/ >>> >> I suggest you upgrade to 2.0.0 as all the known timeout and >> disconnection issues have been fixed. > > you can also put 'option ping-timeout 120' in your volumn definitions, > eg. > > volume whatever > type protocol/client > option transport-type tcp > option remote-host remotehostname > option remote-subvolume brick0 > option ping-timeout 120 > end-volume >Thank You for advice. It work perfect for me. I'm just testing 2.0.0 on 2 boxes as servers with 'replicate translator' and on 100Mb ethernet. One of boxes were getting frozen after about 12-24 hours of bonnie++ tests with various error communicates (cluster as a whole were working). Now with 'option ping-timeout 120' all boxes are running third day of continous bonnie++ test. konrad szeromski
On Thu, May 14, 2009 at 11:57 AM, cnr <cnr.roxx at gmail.com> wrote:> > >> >> > > One of boxes were getting frozen after about 12-24 hours of bonnie++ > tests with various error communicates (cluster as a whole were working). > >Would it be possible to send across the glusterfs log file when you encounter this problem? Thanks, Vijay -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20090514/3cf66f1f/attachment.html>