We’ve got a bunch of new FAS6000 and FAS3000 series clusters in recently and CFO seems to take less than 20 seconds during testing before any real load is placed on them (minimal load during testing).  We’ve seen between 4 and 12 seconds for CF Takeover, and 1 and 16 seconds for CF Giveback.

 

Using loopback adapters for ‘open’ FC ports help as well.

 

 

How long should it take a filer to fail over from one head to another? When I force a failover (cf forcetakeover) from one head the other goes down for minutes. Here’s what I see on the console. This is a new filer with very little traffic going to it and FC is not even set up yet, all NFS/CIFS.

 

array01> cf forcetakeover

cf forcetakeover may lead to data corruption; really force a takeover? yes

cf: forcetakeover initiated by operator

array01> Mon May  5 12:27:48 EST [array01: cf.misc.operatorForcedTakeover:warning]: Cluster monitor: forced takeover initiated by operator

Mon May  5 12:27:48 EST [array01: cf.fsm.takeover.forced:info]: Cluster monitor: takeover attempted after cf forcetakeover command

Mon May  5 12:27:48 EST [array01: cf.fsm.stateTransit:warning]: Cluster monitor: UP --> TAKEOVER

Mon May  5 12:27:48 EST [array01: cf.fm.takeoverStarted:warning]: Cluster monitor: takeover started

Mon May  5 12:27:48 EST [array02/array01: coredump.spare.none:info]: No sparecore disk was found.

Mon May  5 12:27:51 EST [array01: raid.vol.replay.nvram:info]: Performing raid replay on volume(s)

Mon May  5 12:27:51 EST [array01: raid.cksum.replay.summary:info]: Replayed 0 checksum blocks.

Mon May  5 12:27:51 EST [array01: raid.stripe.replay.summary:info]: Replayed 0 stripes.

Mon May  5 12:27:54 EST [array02/array01: wafl.replay.done:info]: WAFL log replay completed, 2 seconds

ifconfig: no such media type <xxx>

         media type options are: <tp> <tp-fd> <100tx> <100tx-fd> <1000fx> <auto> <10g-sr>

ifconfig: Unable to determine primary for interface e0a.

ifconfig: e0a: no such interface

ifconfig: Unable to determine primary for interface e0b.

ifconfig: e0b: no such interface

ifconfig: Unable to determine primary for interface e0c.

ifconfig: e0c: no such interface

ifconfig: Unable to determine primary for interface e0d.

ifconfig: e0d: no such interface

ifconfig: Unable to determine primary for interface e2a.

ifconfig: e2a: no such interface

ifconfig: Unable to determine primary for interface e2b.

ifconfig: e2b: no such interface

add net default: gateway 10.28.17.1: network unreachable

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e0a.

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e0b.

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e0c.

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e0d.

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e2a.

Mon May  5 12:27:55 EST [array02/array01: net.ifconfig.noLocal:error]: ifconfig: Unable to determine primary for interface e2b.

Mon May  5 12:27:55 EST [array02/array01: nis.servers.not.available:error]: NIS server(s) not available.

Mon May  5 12:27:55 EST [array02/array01: cf_takeover:info]: relog syslog Mon May  5 12:26:00 EST [array02: monitor.globalStatus.ok:info]: The system's global status is normal.

Mon May  5 12:27:55 EST [array02/array01: cf_takeover:info]: relog syslog Mon May  5 12:27:47 EST [array02: cf.fsm.takeoverOfPartnerDisabled:notice]: Cluster monitor: takeover of array

There are 68 spare disks; you may want to use the vol or aggr command

to create new volumes or aggregates or add disks to the existing aggregate.

FCP service stopped.

Mon May  5 12:27:55 EST [array01: net.ifconfig.takeoverError:warning]: WARNING: 6 errors detected during network takeover processing WARNING: Some network clients may not be able to access the cluster during takeover

Mon May  5 12:27:55 EST [array01: cf.rsrc.takeoverOpFail:error]: Cluster monitor: takeover during ifconfig_2 failed; takeover continuing...

CIFS partner server is running.

Mon May  5 12:27:55 EST [array01 (takeover): cf.rsrc.transitTime:notice]: Top Takeover transit times wafl_replay=2383 {replay_log=2353, mark_replaying=29}, raid=832, rc=410 {hostname=51, ifconfig=46, options=23, options=14, options=10, options=9, ifconfig=1, ifconfig=1, ifconfig=1, route=1}, wafl=405, registry_postrc_phase1=227, raid_replay=179, registry_prerc=115, wafl_sync=74, fmdisk_reserve=70, cifs=70

Mon May  5 12:27:55 EST [array01 (takeover): cf.fm.takeoverComplete:warning]: Cluster monitor: takeover completed

Mon May  5 12:27:55 EST [array01 (takeover): cf.fm.takeoverDuration:warning]: Cluster monitor: takeover duration time is 7 seconds

Mon May  5 12:27:58 EST [array02/array01: asup.smtp.host:info]: Autosupport cannot connect to host smtp.danahermail.com (Network comm problem) for message: REBOOT (CLUSTER TAKEOVER)

Mon May  5 12:27:58 EST [array02/array01: asup.smtp.unreach:error]: Autosupport mail was not sent because the system cannot reach any of the mail hosts from the autosupport.mailhost option. (REBOOT (CLUSTER TAKEOVER))

Mon May  5 12:28:00 EST [array01 (takeover): monitor.globalStatus.critical:CRITICAL]: This node has taken over array02.

Mon May  5 12:28:00 EST [array02/array01: monitor.globalStatus.critical:CRITICAL]: array01 has taken over this node.

Mon May  5 12:28:05 EST [array02/array01: nbt.nbns.registrationComplete:info]: NBT: All CIFS name registrations have completed for the partner server.

Mon May  5 12:28:07 EST [array01 (takeover): asup.post.sent:notice]: Cluster Notification message posted to IBM: Cluster Notification from array01 (CLUSTER TAKEOVER COMPLETE MANUAL) INFO

Mon May  5 12:32:07 EST [array02/array01: asup.smtp.host:info]: Autosupport cannot connect to host smtp.danahermail.com (Network comm problem) for message: REBOOT (CLUSTER TAKEOVER)

Mon May  5 12:32:07 EST [array02/array01: asup.smtp.unreach:error]: Autosupport mail was not sent because the system cannot reach any of the mail hosts from the autosupport.mailhost option. (REBOOT (CLUSTER TAKEOVER))

 

Here is the ifconfig –a from the node that stayed up:

array01(takeover)> ifconfig -a

e0a: flags=948043<UP,BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 02:a0:98:08:22:b7 (auto-1000t-fd-up) flowcontrol full

        trunked lan0

e0b: flags=108042<BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 00:a0:98:08:22:b6 (auto-unknown-cfg_down) flowcontrol full

e0c: flags=948043<UP,BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 02:a0:98:08:22:b7 (auto-1000t-fd-up) flowcontrol full

        trunked lan0

e0d: flags=108042<BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 00:a0:98:08:22:b4 (auto-unknown-cfg_down) flowcontrol full

e2a: flags=108042<BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 00:07:43:05:16:98 (auto-10g_sr-fd-cfg_down) flowcontrol full

e2b: flags=108042<BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        ether 00:07:43:05:16:99 (auto-10g_sr-fd-cfg_down) flowcontrol full

lo: flags=19e8049<UP,LOOPBACK,RUNNING,MULTICAST,MULTIHOST,PARTNER_UP,TCPCKSUM> mtu 8160

        inet 127.0.0.1 netmask 0xff000000 broadcast 127.0.0.1

        ether 00:00:00:00:00:00 (VIA Provider)

lan0: flags=948043<UP,BROADCAST,RUNNING,MULTICAST,TCPCKSUM> mtu 1500

        inet 10.28.17.213 netmask 0xffffff00 broadcast 10.28.17.255

        partner lan0 (not in use)

        ether 02:a0:98:08:22:b7 (Enabled virtual interface)