On Thu, Feb 14, 2008 at 12:10 PM, Keshetti Mahesh <keshetti85-student@xxxxxxxxxxx> wrote: > > Can you send me /proc/partitions file of master and slave nodes after > > booting into GlusterHPC? > > ==================================================== > ----> On server (n1), > > # cat /proc/partitions > > i) When booted with the GlusterHPC, > major minor #blocks name > 22 0 80418240 hdc > 22 1 104391 hdc1 > 22 2 19077187 hdc2 > 22 3 49145715 hdc3 > 22 4 1 hdc4 > 22 5 8385898 hdc5 > 22 6 1090071 hdc6 > > ii) when booted with OS (FC7), > major minor #blocks name > 8 0 80418240 sda > 8 1 104391 sda1 > 8 2 19077187 sda2 > 8 3 49145715 sda3 > 8 4 1 sda4 > 8 5 8385898 sda5 > 8 6 1090071 sda6 > > ==================================================== > > ----> On client (n2) : there is no OS installed on client > > # cat /proc/partitions > > (When booted with GlusterHPC) > > major minor #blocks name > 3 0 80418240 hda > > ==================================================== > > If you need any more information please feel free to ask. > > -Mahesh > Just now I got a hint on whats happening in the partition naming of the Gluster. In my server (n1), SATA disk is located in the IDE channel #1 as master whereas the the same capacity SATA disk is located in IDE channel #0 as master in the client node (n2). This difference in the location of hard disk is letting the Gluster to name the hard disk drive devices differently on the two nodes. i.e. SATA disk is named as hdc in server and as hda in client. Due to this difference the client is generating errors like "Can' read from the /dev/hdc3: there is no devic ewith that name" and aborting. If it is the actual reason behind the problem I have reported it is really a bug in Gluster. -Mahesh
