Jason Haar wrote: > To reiterate: I have two bricks (both /home/export) via 2 servers, > mounted on two clients as /mnt/glusterfs. I then copied data (on one > client) into /mnt/glusterfs and saw the data start showing up under > /home/export on both servers. However, after killing the copy, I have > inconsistent files under /home/export. I then ran > "find /mnt/glusterfs/ -type f -print0 | xargs -0 head -c1 > /dev/null" > to force glusterfs to read every file, and that didn't change the state > of /home/export on the two bricks. Ah, I've just noticed that running > "du /mnt/glusterfs" on both clients also produces inconsistent output - > that certainly shouldn't be happening? > > ...and it's repeatable for me. I just umounted the two clients, killed the server daemons and restarted glusterfs again, did a copy and got similar results. First I copied a single file - and that worked fine. Then I started a "rsync -a /etc/ /mnt/glusterfs/newDir" - and killed it off. Then I did a "find /mnt/glusterfs/newDir -exec cat..." to ensure both clients should be in sync - and they aren't??? I see a whole bunch of these in the glusterfsd.log file of the server with the *larger* number of files 2008-10-11 09:25:40 W [posix.c:912:posix_utimens] brick: utimes on /etc/rc.local: No such file or directory I think they are ignorable - they appears to all refer to symlinks - but I'll mention them anyway. -- Cheers Jason Haar Information Security Manager, Trimble Navigation Ltd. Phone: +64 3 9635 377 Fax: +64 3 9635 417 PGP Fingerprint: 7A2E 0407 C9A6 CAF6 2B9F 8422 C063 5EBB FE1D 66D1
