[Gluster-devel] What does it mean (unify bug ?) ?

[Anton Khalikov]

Hello guys

I use glusterfs for a distributed backup storage. At the moment there
are 2 servers with 300 Gb ext3 fs on lvm2 joined into single namespace
using UNIFY.

Sometimes (in fact very often) I get the following error:

# head /backup/hosting/*****/******/log/access.log
head: cannot open `/backup/hosting/*****/*****/log/access.log' for
reading: Input/output error

In the same moment glusterfs.log gets the following:

2008-06-02 10:30:54 E [unify.c:850:unify_open]
backup-unify: /hosting/*****/******/log/access.log: entry_count is 3
2008-06-02 10:30:54 E [unify.c:853:unify_open]
backup-unify: /hosting/*****/******/log/access.log: found on bbrick2
2008-06-02 10:30:54 E [unify.c:853:unify_open]
backup-unify: /hosting/*****/******/log/access.log: found on bbrick1
2008-06-02 10:30:54 E [unify.c:853:unify_open]
backup-unify: /hosting/*****/******/log/access.log: found on afr-ns
2008-06-02 10:30:54 E [fuse-bridge.c:675:fuse_fd_cbk] glusterfs-fuse:
34127360: /hosting/*****/******/log/access.log => -1 (5)

I understand that it means that the file exists on the both servers
which is wrong but how is this possible if I created a fresh empty
storage 2 days ago ?

glusterfs 1.3.8pre1 built on Feb 26 2008 18:21:47
Repository revision: glusterfs--mainline--2.5--patch-676

Installed from lmello's debian packages on debian etch system.

--
Best regards
Anton Khalikov