Hi everydoby, i'm new in this ng :=)
I'm also new to cluster and i have little questions.
I have 2 node (identical, HP DL380) and a san MSA1500, redundant path FC,2
network card for private and public.SQL 2000 clustered.
These are my tests:
1. i try to unplug the cable (private connection, hbeat) but i don't know
what must happen. Is the main node that must remain up and running?
2.when i unplug the public cable all the group move to nodeB except the
group "Storage" where i put three physical disk, why? The physical disk have
no dependencies...i don't know this behavior :-\
3. i try to unplug the FC connection fro the server and the groups move from
nodeA to nodeB. When i plug into the FC connection the group remains on the
nodeB instead to move on nodeA. The failback setting are correct, because if
i power down nodeA or nodeB the "Cluster cycle" work correct: from A to B
and when A is online the resouce came back again to nodeA. It's seems "don't
hear" the change of state of the FC link...
There are some techicnal limitation that i don't know or there are some
setting to do?
Thanks a lot in advance!
ciao
mn

Comments Inline

[quoted text, click to view]

That's OK, I have been here for a while. Everybody starts at the beginning.

[quoted text, click to view]
Probably nothing if you have configured everything correctly. The cluster
should use the public NICs as a backup communications path. This is
determined by the cluster configuration network settings. Since no resource
dependencies were lost AND the cluster nodes still have a communication
path, nothing changes.

[quoted text, click to view]
This is normal. When you unplugged the public NIC, all IP addresses that
depend on that physical resource went offline. When each resource went
offline, the cluster moved the recource group it was part of to another host
in order to try and bring it online. Since the group "Storage" had no
network dependencies, it didn't go anywhere.


[quoted text, click to view]
It may take a reboot to restore the disk connection via the Fibre Channel
connection. You may be able to rescan using the Disk Management tool and
get the system to see the disks again. Either way, a reboot after such a
disconnect is not the worst idea in the world.

[quoted text, click to view]
Glad to help

Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator

Thank u very much for your help Geoff!
My godness...how many thing i have to learn... =:-)
Your explanation is perfectly clear for me, but about one thing im' not
sure:
The group "Storage" is composed by 3 physical disk only, is correct? In
other words, is correct that this group don't move from one node to the
other or i have to put in this group a dependencie (network, ip...)?
For the other points, they are very clear for me now.
Have u got a link for this type of troubleshooting or a book to suggest?
Thanks a lot again!
Ciao
mn

To use these disks for SQL, you will need to move each disk to the resource
group for whichever SQL instance you want to use it for. You will then have
to modify the dependencies of the SQL Server service to be dependent on that
disk. The SQL Service will have to be offline to make this change. Read
BOL on FAILOVER CLUSTERING for more details.

Geoff N. Hiten
Microsoft SQL Server MVP

[quoted text, click to view]

"Geoff N. Hiten" <sqlcraftsman@gmail.com> ha scritto nel messaggio
news:u47NwjySFHA.2256@tk2msftngp13.phx.gbl...
[quoted text, click to view]

Ok, now i know!
It's completely clear! :=)
Thanks a lot again!
Ciao
mn