K 10
svn:author
V 7
asomers
K 8
svn:date
V 27
2013-10-14T21:59:31.240078Z
K 7
svn:log
V 1536
zfsd should activate a spare that is newly added to an already-degraded pool

	cddl/sbin/zfsd/case_file.cc
	cddl/sbin/zfsd/case_file.h
	cddl/sbin/zfsd/dev_ctl_event.cc
		When a config_sync event arrives, reevaluate all open
		casefiles on the same pool.

	cddl/sbin/zfsd/case_file.cc
		When evaluating a config_sync event, try to activate spares
		if our pool is unhealthy.  This is necessary because we
		don't get any more specific event when a spare is added to a
		pool.  This required adding the CaseFileReEvaluator functor
		and the CaseFile::ReEvaluateByGuid method.

	cddl/sbin/zfsd/case_file.cc
		When processing a statechange event, refresh the vdev state.
		This is necessary because we could be reevaluating an
		already-existing casefile that was opened before the pool
		became degraded.  Also, always consume the event, even if we
		couldn't activate the spare.  It's no longer necessary to
		queue the statechange event since we can now reevaluate
		CaseFiles on a config_sync event.

	cddl/sbin/zfsd/case_file.cc
		In CloseIfSolved(), don't close casefiles for DEGRADED or
		FAULTED vdevs since we now have the ability to fix them.

	cddl/sbin/zfsd/dev_ctl_event.cc
	cddl/sbin/zfsd/dev_ctl_event.h
	cddl/sbin/zfsd/zfsd.cc
		Redelegated responsibility for queueing events from
		DevCtlEvent(and its child classes) to ZfsDaemon.

	cddl/sbin/zfsd/zfsd.cc
		In the SIGINFO handler, print unconsumed events as well as
		saved casefiles.

Submitted by:	alans
Approved by:	ken (mentor)
Sponsored by:	Spectra Logic Corporation

END