[Ocfs2-users] Re: ocfs2_broadcast_vote ERROR: status = -75 on Fedora 7 kernel-2.6.22.5-76.fc7

[davide rossetti]

On 9/19/07, davide rossetti <davide.rossetti at gmail.com> wrote:
> dear all,
> I'm a bit depressed as of now... I upgraded my OCFS2 cluster to the
> latest F7 kernel 2.6.22.5-76.fc7 and I can't mount my filesystem on
> two PC at the same time, one is 32bit and the other is 64.... any idea
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_broadcast_vote:434
> ERROR: status = -75
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_do_request_vote:504
> ERROR: status = -75
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_mount_volume:1117
> ERROR: status = -75
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_broadcast_vote:434
> ERROR: status = -75
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_do_request_vote:504
> ERROR: status = -75
> Sep 19 20:15:29 theboss kernel: (29057,2):ocfs2_dismount_volume:1179
> ERROR: status = -75
> Sep 19 20:15:33 theboss kernel: ocfs2: Unmounting device (8,33) on (node 3)
> Sep 19 20:15:35 theboss kernel: o2net: no longer connected to node
> rack1 (num 1) at 10.0.2.21:7777

investigating a little bit more, seems like (but not sure) EOVERFLOW
can be thrown only in ocfs2/cluster/tcp.c

			if (sc->sc_page_off == sizeof(struct o2net_msg)) {
				hdr = page_address(sc->sc_page);
				if (be16_to_cpu(hdr->data_len) >
				    O2NET_MAX_PAYLOAD_BYTES)
					ret = -EOVERFLOW;
			}

I guess there could be some 32/64 cleanness problem somewhere...

I'd like to debug it a bit more but I'm not able to activate enough
debugging even after enabling it:
[root at theboss ~]# /sbin/debugfs.ocfs2 -l  /dev/sdc1
KTHREAD off
NOTICE allow
ERROR allow
EXPORT off
QUORUM off
CONN off
DCACHE off
VOTE allow
INODE off
NAMEI off
UPTODATE off
BH_IO off
DLM_GLUE off
EXTENT_MAP off
FILE_IO off
SUPER allow
DISK_ALLOC off
JOURNAL off
AIO off
DLM_RECOVERY off
DLM_MASTER off
DLM_THREAD off
DLM_DOMAIN off
DLM off
DLMFS off
HB_BIO off
HEARTBEAT off
SOCKET off
MSG allow
TCP allow
EXIT deny
ENTRY deny
.. there must be some 'turn-on debugging log' master switch somewhere...

-- 
davide.rossetti at gmail.com ICQ:290677265 SKYPE:d.rossetti