[Ocfs2-users] mounted.ocfs2 -f shows Unknown

[David Johle]

So to doublecheck that, I get this on the unknown node (grate):


# cd /sys/kernel/config/cluster/live2/node
# for x in `find . -type f` ; do echo -n "$x:" ; cat $x; done
./grate/local:1
./grate/ipv4_address:1.2.3.47
./grate/ipv4_port:7777
./grate/num:3
./whip/local:0
./whip/ipv4_address:1.2.3.42
./whip/ipv4_port:7777
./whip/num:2
./pig/local:0
./pig/ipv4_address:1.2.3.41
./pig/ipv4_port:7777
./pig/num:1


Did this on the other node that is up (whip) and the results were the 
same except for the "local" value being a 0 for grate and a 1 for 
whip like it should.




At 04:06 PM 1/12/2010, Sunil Mushran wrote:
>mounted.ocfs2 looks up 
>/sys/kernel/config/cluster/<clustername>/node/<nodename>/num
>to match the node number with the node name. If it does not find it, 
>it prints unknown.
>
>
>David Johle wrote:
>># debugfs.ocfs2 -R "slotmap" /dev/dm-6
>>Slot# Node#
>>0 3
>>1 2
>>
>># debugfs.ocfs2 -R "hb" /dev/dm-6
>>node: node seq generation checksum
>>1: 1 000000004b3949ce 0000000000000000 fb8894ff
>>2: 2 000000004b4cd7d0 e869dccb731433ed 805d51fc
>>3: 3 000000004b4cd796 2713837530adc63c 7c4bb936
>>
>># debugfs.ocfs2 -R "hb" /dev/dm-6
>>node: node seq generation checksum
>>1: 1 000000004b3949ce 0000000000000000 fb8894ff
>>2: 2 000000004b4cd7d8 e869dccb731433ed 03cb88bc
>>3: 3 000000004b4cd79e 2713837530adc63c ffdd6076
>>
>>At 06:46 PM 1/11/2010, Sunil Mushran wrote:
>>>Email me the outputs of the following:
>>>
>>># debugfs.ocfs2 -R "slotmap" /dev/dm-6
>>>
>>># debugfs.ocfs2 -R "hb" /dev/dm-6
>>>wait 10 seconds.
>>># debugfs.ocfs2 -R "hb" /dev/dm-6
>>>
>>>
>>>David Johle wrote:
>>>>I was setting up a new 3 node cluster of systems and just so 
>>>>happened to have only 2 of the nodes online after a kernel 
>>>>update. Those nodes were numbers 2 & 3 in the cluster.conf.
>>>>
>>>>It seems that the issue referenced here...
>>>>http://kr.forums.oracle.com/forums/thread.jspa?threadID=1005937
>>>>
>>>>...is still around. Not so much that I have configured a node 
>>>>number too high (they are 1,2,3 and node count is 3), but rather 
>>>>there is a node that joined the domain who's number is higher 
>>>>than the total number of nodes joined thus far.
>>>>
>>>>----sample output----
>>>>
>>>># mounted.ocfs2 -f
>>>>
>>>>Device FS Nodes
>>>>/dev/dm-6 ocfs2 Unknown, whip
>>>>
>>>>----sample output----
>>>>
>>>>
>>>>
>>>>Has anyone else run across this, or can anyone else confirm this 
>>>>happens and that it's not just something configured wrong on my 
>>>>system (I've double checked many times).
>>>>
>>>>It seems to only be a cosmetic thing for this command, everything 
>>>>works fine as far as filesystem access and whatnot.