Can the cf giveback -f command be disruptive?
Applies to
Data ONTAP 8.2 7-Mode
Answer
Yes it is disruptive.
- When performing a
cf giveback
, the system performs a number of checks to prevent the giveback from being disruptive. This can include checking for failed disks, active CIFS sessions, active backups, and the HA partner's status. - The
cf giveback -f
command will bypass any of these checks. This command will only be canceled or vetoed by the system if there is a risk of data loss. Thecf giveback -f
command will allow a giveback in situations where an interruption in service is possible. - For example, the
cf giveback -f
command will perform a giveback when the partner is not ready for a giveback.
storage1(takeover)> cf status
storage1 has taken over storage2.
storage1(takeover)> cf giveback
Partner not waiting for giveback, giveback cancelled.
To perform a giveback without checking for partner readiness, either set the cf.giveback.check.partner option to off before performing a cf giveback again or Perform a cf giveback -f.
The first choice disables checking for all future cf giveback, until it is turned back to on. The second choice is good for this giveback only.
storage1(takeover)> cf giveback -f
Fri Jan 31 19:29:50 EST [storage1:cf.misc.operatorGiveback:info]: Failover monitor: giveback -f initiated by operator
Fri Jan 31 19:29:50 EST [storage1:cf.misc.givebackPartnerNotReady:notice]: Giveback is being initiated before the partner is in the waiting-for-giveback state. The partner is in the SF_UNKNOWN state.
Fri Jan 31 19:29:50 EST [storage1:cf.fm.givebackStarted:notice]: Failover monitor: giveback started.
Fri Jan 31 19:29:53 EST [storage1:cf.fm.gi2220cl1vebackComplete:notice]: Failover monitor: giveback completed
Fri Jan 31 19:29:53 EST [storage1:cf.fsm.stateTransit:info]: Failover monitor: TAKEOVER --> UP
Fri Jan 31 19:29:58 EST [storage1:cf.fsm.partnerNotResponding:notice]: Failover monitor: partner not responding
Fri Jan 31 19:30:07 EST [storage1:ha.takeoverImpNotDef:warning]: Takeover of the partner node is impossible due to reason waiting for partner to recover.
- The
cf giveback -f
command could also result in interruptions where the HA partner is waiting for a giveback.
Perform the following steps instead of running the cf giveback -f
command:
cf status
:
Verify that the HA partner is available and waiting for a giveback. Connecting to the console of the HA partner should display the following message if the system is ready for a giveback:
Waiting for giveback...(Press Ctrl-C to abort wait)
cf giveback
:
Review the messages that follow. If the giveback is cancelled, messages are reported to indicate why the giveback was cancelled.- Address each of these items individually prior to performing the giveback.
- Attempt
cf giveback
again. - Review the messages for any indication of services or situations that prevent the giveback.
Run the cf giveback -f
command only if the above steps have been performed and only under the direction of NetApp Support.
Additional Information
BUG 553014