Hi, I'm faced with a very strange behaviour:
Centos 6.5 server, Hardware ISCSI HBA from Emulex OneConnect, most recent drivers and firmware from emulex installed.
Directly attached a 10G ISCSI Storage from QSan. Two Raid volumes, Raid 5, 8 Disks each at 2 TB so 14 TB each logic raid volume.
Both volumes are logged in and usable as sdb and sdc to the server.
Formatted with ext4 -m 0 -v /dev/sd...
Now I copied about 600 GB from an old server to one of the new volumes. No problems at all.
But moving other data from the old server to the second volume shows the error in the logs:
EXT4-fs error (device sdc): ext4_mb_generate_buddy: EXT4-fs: group 2592: 0 blocks in bitmap, 24544 in gd
group grows by 16.
But the more strange thing I realised today is, that all the data on the first volume (on sdb) shows up on sdc as well + the new data too.
Now I'm totally confused and currently there are no more messages in the system log.
The files on sdc don't show up on sdb!
If I do e.g. a md5sum on the files on both volumes they are the same as is the contents of e.g. some textfiles I checked.
May be someone has an idea what might cause
a) the error and b) why dose the data from sdb also shows up on sdc?
b) also occurred also with the original redhat/centos driver for the ISCSI HBA.
May be I have to use partitions and not the whole device or LVM, ... xfs?
Other storages from Qsan dont show that problem and other storages with bigger filesystems or e.g. xfs are o.k. too.
confused regards . Götz