fair bloom
Hi mates I have 2 netapp fas2750, I will do a shutdown today like in one hour because client is moving to another location, HA is enabled, should I disable after entering the suppression maintenance mode? to avoid LIF move between ports and failover stuff? things to keep in mind? thanks, I will follow this documentation @grand cape on NetApp
grand capeBOT
Hi mates I have 2 netapp fas2750, I will do a shutdown today like in one hour be...
Sorry, this request does not comply with MEE6 content policy. Modify it and try again.
urban wave
that KB follows everything you need to know
are they 2 seperate clusters or a single 4 node cluster?
frosty badger
For those wondering, the easy thing to do
set diag
cluster show (look for node with epsilon, shut it down last or modify the node to remove epsilon and move it to the node you are on)
Halt -node node-04 -skip-lif-migration true -inhibit-takeover true -ignore-quorum-warning true
Repeat for the nodes and do the one with epsilon last
When you power on, the nodes will just “return to service”
radiant delta
...or just do set -confirmations off followed by node halt * -inhibit-takeover true -skip-lif-migration-before-shutdown true
frosty badger
That doesn’t always work. Even with just ha, sometimes the ignore-quorum-warning if necessary. Hence find the epsilon and shutdown last
Had it happen yesterday
radiant delta
right, you need it for smsync for example I think... and probably smas/smbc. it doesn't hurt to add this as well though. If you shut down the nodes with a wildcard you don't need to reconnect your SSH session after every node shuts down (Murphy's Law: your cluster mgmt LIF is always on the node you're about to shut down 😄 )
fair bloom
ty I followed that KB article
haughty citrus
...or just do `set -confirmations off` followed by `node halt * -inhibit-takeove...
confirmations off takes out a lot of warnings and error handling out. Would highly recommend to never use it.
radiant delta
confirmations off takes out a lot of warnings and error handling out. Would high...
it only persists for as long as the SSH session is there, and in this case you're halting the nodes anyway, so there is hardly anything else it could affect 🙂
but even in other cases, "never use it" is probably taking it a bit too far. For example try deleting a couple dozen snapshots on a couple dozen volumes each with confirmations on 😄
haughty citrus
but even in other cases, "never use it" is probably taking it a bit too far. For...
until you notice you deleted the wrong one as you did not doublecheck 😉
radiant delta
yeah, well, that's what we call a "learning experience" then I guess 😂
fresh eagle
one can avoid the confirmations on a lot of operations with -force ... again, YMMV
radiant delta
one can avoid the confirmations on a lot of operations with -force ... again, YM...
not on "node halt"
fresh eagle
not on "node halt"
was directed more at snapshot deletion
radiant delta
was directed more at snapshot deletion
that's about the only command where they are (almost) equivalent. For volumes, for example, I'd argue that using -force is actually more dangerous than using set -confirmations off since "force" skips the recovery queue while "confirmations off" doesn't
fresh eagle
yes, the caveat of YMMV ... removing confirmations has to be re-enabled whereas -force is just for the one command...