On 18/07/11 02:05, Dan Bretherton wrote: > ---------------------------------------------------------------------- >> Message: 1 >> Date: Fri, 8 Jul 2011 15:54:19 -0700 >> From: Vikas Gorur<vikas at gluster.com> >> Subject: GFID mismatches and tools to fix them >> To: gluster-users at gluster.org >> Cc: gluster-devel at gluster.com >> Message-ID: >> <CAHVddbPm1YARBzMevK6hyN8U9yt-ayETg65i__SFMMNp0NvmMQ at mail.gmail.com> >> Content-Type: text/plain; charset="iso-8859-1" >> >> Hi everyone, >> >> As some of you may know, the presence of files/directories which have >> different GFID's on different backends can cause the GlusterFS client to >> throw up errors, or even hang. Among others, we've had users of Enomaly >> affected by this issue. >> >> A little background on GFIDs: >> >> Each file/directory on a Gluster volume has a unique 128-bit number >> associated with it called the GFID. This is true regardless of Gluster >> configuration (distribute or distribute/replicate). One inode, one GFID. >> >> The GFID is stored on the backend as the value of the extended attribute >> "trusted.gfid". Under normal circumstances, the value of this >> attribute is >> the same on all the backend bricks. However, certain conditions can >> cause >> the value on one or more of the bricks to differ from that on the other >> bricks. This causes the GlusterFS client to become confused and throw up >> errors. This applies to both the 3.1.5 and 3.2.1 versions of the >> filesystem, >> and previous versions in those series. >> >> In a future bugfix release GlusterFS will fix this issue >> automatically when >> it detects it. Until then, if you encounter this problem, please use the >> following set of tools to manually fix it on the backends: >> >> https://github.com/vikasgorur/gfid >> >> The repository contains the tools as well as a README that explains >> how to >> use them. >> >> Your questions and comments are welcome. >> > Dear Vikas- > Thanks for provding these tools. Unfortunately I think I have found a > problem with the procedure outlined in the README - I don't think it > works for files with names containing the colon character. I still > have a lot of gfid errors in my logs after running the gfid tools on > one volume, and all the filenames have one or more ':' characters. > There are 1677 files still affected with "gfid different" so I don't > think it can be a coincidence. > > Regards > -Dan. I had a closer look at this. It is the output of gfid-mismatch causing the problem; paths are shown with a trailing colon as in GlusterFS log files. The "cut -f1 -d:" to extract the paths obviously removes all the colons. I'm sure there is an easy way to remove the trailing ':' from filenames but I can't think of one off hand (and it is 3:30AM). -Dan.
