Change in vdsm[master]: Support for GLUSTERFS_DOMAIN

deepakcs at linux.vnet.ibm.com deepakcs at linux.vnet.ibm.com
Wed Aug 29 11:29:29 UTC 2012


Deepak C Shetty has posted comments on this change.

Change subject: Support for GLUSTERFS_DOMAIN
......................................................................


Patch Set 5: No score

(4 inline comments)

@Saggi,
   Per our IRC chat, I discussed with the gluster folks and it seems I can live w/o worrying about transport and port as inputs, thus there shouldn't be the need to persist them, and hence no need of REMOTE_INFO MD key. I will send a updated patch soon.

....................................................
Commit Message
Line 18: 
Line 19: This patch attempts to re-use nfsSD core logic to support
Line 20: domain of type GLUSTERFS_DOMAIN.
Line 21: 
Line 22: v2:
Done
Line 23:     Using different key drive['vmVolInfo'] instead of 'path'
Line 24:     Addressed most of federico's comments
Line 25: v3:
Line 26:     Addressed federico's comments


....................................................
File vdsm/storage/blockSD.py
Line 469:         t = time.time()
Line 470:         misc.readfile(lvm.lvPath(self.sdUUID, sd.METADATA), 4096)
Line 471:         return time.time() - t
Line 472: 
Line 473: 
In discussion with federico, it was decided to unify the produceVolume, hence this was removed. Its related because for the newly introduced glusterVolume class, won't get instantiated, if called via produceVolume, as produceVolume was not asking getVolumeClass, it was hardcoded. So when I changed produceVolume to work via getVolumeClass, it made sense to unify them too.
Line 474:     def getVolumeClass(self):
Line 475:         """
Line 476:         Return a type specific volume generator object
Line 477:         """


....................................................
File vdsm/storage/hsm.py
Line 124:     # FCP domain shouldn't even be on the list but VDSM use to just
Line 125:     # accept this type as iscsi so we are stuck with it
Line 126:     sd.FCP_DOMAIN: 'iscsi',
Line 127:     sd.POSIXFS_DOMAIN: 'posixfs',
Line 128:     sd.GLUSTERFS_DOMAIN: 'posixfs' }
Because I want to re-use the nfsSD core logic, that is done in PosixFS.
Line 129: 
Line 130: def _BCInitiatorNameResolve(ifaceName):
Line 131:     if not ifaceName:
Line 132:         return iscsi.IscsiInterface('default')


....................................................
File vdsm/storage/nfsSD.py
Line 122:     def getVolumeClass(self):
Line 123:         """
Line 124:         Return a type specific volume generator object
Line 125:         """
Line 126:         if self.getMetaParam(sd.DMDK_TYPE) == sd.GLUSTERFS_DOMAIN:
Accessor ?
Line 127:             return glusterVolume.GlusterVolume
Line 128:         else:
Line 129:             return fileSD.FileStorageDomain.getVolumeClass(self)
Line 130: 


--
To view, visit http://gerrit.ovirt.org/6856
To unsubscribe, visit http://gerrit.ovirt.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I9ac37da88625f20d148beaf53bb6371c15b33ad7
Gerrit-PatchSet: 5
Gerrit-Project: vdsm
Gerrit-Branch: master
Gerrit-Owner: Deepak C Shetty <deepakcs at linux.vnet.ibm.com>
Gerrit-Reviewer: Adam Litke <agl at us.ibm.com>
Gerrit-Reviewer: Dan Kenigsberg <danken at redhat.com>
Gerrit-Reviewer: Deepak C Shetty <deepakcs at linux.vnet.ibm.com>
Gerrit-Reviewer: Federico Simoncelli <fsimonce at redhat.com>
Gerrit-Reviewer: Royce Lv <lvroyce at linux.vnet.ibm.com>
Gerrit-Reviewer: Saggi Mizrahi <smizrahi at redhat.com>
Gerrit-Reviewer: Shu Ming <shuming at linux.vnet.ibm.com>
Gerrit-Reviewer: oVirt Jenkins CI Server


More information about the vdsm-patches mailing list