I can comment on the IOPS part. The IOPS are currently only relevant when using managed storage (which is currently only applicable at the zone level). That being the case, no cluster (or local) storage takes advantage of counting IOPS that have been handed out.
Of course this logic might be expanded in the future, but that's how it works today. On Tue, Jun 17, 2014 at 2:47 PM, Prachi Damle <prachi.da...@citrix.com> wrote: > >Why ZoneWideStoragePoolAllocator implements "filter" method (and doesn't > call that of base class) rather than just using that of base class? > >ZoneWideStoragePoolAllocator#filter method seems doesn't care "avoid" > parameter, doesn't skip iSCSI type storage pool even if a volume is a > RootDisk. > > The first question seems like a bug to me. > > I am not sure about the IOPS > > Prachi > > -----Original Message----- > From: ynoj...@ynojima.net [mailto:ynoj...@ynojima.net] On Behalf Of > Yoshikazu Nojima > Sent: Tuesday, June 17, 2014 1:07 PM > To: dev@cloudstack.apache.org > Subject: ZoneWideStoragePoolAllocator#filter seems weird for me > > Hi, > > ZoneWideStoragePoolAllocator#filter seems weird for me. > > ZoneWideStoragePoolAllocator and ClusterScopeStoragePoolAllocator are > allocator classes to select storage pool. > They extends AbstractStoragePoolAllocator class, which provides "filter" > method to exclude unavailable storage pools. > > Why ZoneWideStoragePoolAllocator implements "filter" method (and doesn't > call that of base class) rather than just using that of base class? > > ZoneWideStoragePoolAllocator#filter method seems doesn't care "avoid" > parameter, doesn't skip iSCSI type storage pool even if a volume is a > RootDisk. > (These functions are implemented in > AbstractStoragePoolAllocator#filter method, which used by > ClusterScopeStoragePoolAllocator.) > On the other hand, AbstractStoragePoolAllocator#filter doesn't call > storageMgr.storagePoolHasEnoughIops, so a cluster wide primary storage > would be allocated more volumes than its designated IOPS capacity. > > Is there any difference between a zone wide primary storage and a cluster > wide primary storage except its scope? > If it is a bug, I'll fix it. > > Regards, > Noji > -- *Mike Tutkowski* *Senior CloudStack Developer, SolidFire Inc.* e: mike.tutkow...@solidfire.com o: 303.746.7302 Advancing the way the world uses the cloud <http://solidfire.com/solution/overview/?video=play>*™*
