July 2011 - vdsm-patches - Fedora Mailing-Lists

[ABANDONED PATCH] [WIP] BZ#719330 - No need to spew unmounted masters to the log. (via gerrit-bot)

by smizrahi＠redhat.com

Patch was abandoned by Saggi Mizrahi (smizrahi(a)redhat.com) because: Bug is non issue and refactoring can wait You can review this change at: http://gerrit.usersys.redhat.com/683

12 years, 11 months

1
0
0 / 0

[NEW PATCH] [WIP] BZ#719330 - No need to spew unmounted masters to the log. (via gerrit-bot)

by smizrahi＠redhat.com

New patch submitted by Saggi Mizrahi (smizrahi(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/683 commit d334c99c34d3a922a89ff75280e927235b0984db Author: Saggi Mizrahi <smizrahi(a)redhat.com> Date: Mon Jul 11 12:15:48 2011 +0300 [WIP] BZ#719330 - No need to spew unmounted masters to the log. Change-Id: Ib0044075f10171bd37613e17b46b28cd7e81dff7 diff --git a/vdsm/storage/hsm.py b/vdsm/storage/hsm.py index b0d5d21..623b09f 100644 --- a/vdsm/storage/hsm.py +++ b/vdsm/storage/hsm.py @@ -607,7 +607,7 @@ class HSM: for sdUUID in domList: vars.task.getExclusiveLock(STORAGE, sdUUID) - self.spm.detachAllDomains(pool) + self.spm.destroyPool(pool) return self._disconnectPool(pool, hostID, scsiKey, remove=True) diff --git a/vdsm/storage/sp.py b/vdsm/storage/sp.py index 8aed3e6..e9f5c90 100644 --- a/vdsm/storage/sp.py +++ b/vdsm/storage/sp.py @@ -595,31 +595,11 @@ class StoragePool: self.log.error("Unexpected error", exc_info=True) - def __unmountLastMaster(self, sdUUID): - curmsd = SDF.produce(sdUUID) - # Check if it's last domain and allow it detaching - dl = self.getDomains(activeOnly=True) - domList = dl.keys() - if curmsd.sdUUID in domList: - domList.remove(curmsd.sdUUID) - for item in domList: - domain = SDF.produce(item) - if domain.isData(): - # Failure, we have at least one more data domain - # in the pool and one which can become 'master' - raise se.StoragePoolHasPotentialMaster(item) - curmsd.unmountMaster() - def masterMigrate(self, sdUUID, msdUUID, masterVersion): self.log.info("sdUUID=%s spUUID=%s msdUUID=%s", sdUUID, self.spUUID, msdUUID) - # Check if we are migrating to or just unmounting last master - if msdUUID != sd.BLANK_UUID: - self.__masterMigrate(sdUUID, msdUUID, masterVersion) - return False # not last master - - self.__unmountLastMaster(sdUUID) - return True # last master + self.__masterMigrate(sdUUID, msdUUID, masterVersion) + return False # not last master def attachSD(self, sdUUID): """ @@ -651,6 +631,21 @@ class StoragePool: dom.releaseClusterLock() self.updateMonitoringThreads() + def destroy(self): + curmsd = self.getMasterDomain() + # Check if there is only 1 domain and allow it detaching + domList = self.getDomains() + if len(domList) > 1: + raise se.StoragePoolDestroyingError() + + with self._metadata.transaction(): + spUUID = self.spUUID + self._metadata[PMDK_DOMAINS] = {} + self._metadata[PMDK_MASTER_VER] = 0 + curmsd.detachSD(spUUID) + + curmsd.unmountMaster() + def forcedDetachSD(self, sdUUID): self.log.warn("Force detaching domain `%s`", sdUUID) diff --git a/vdsm/storage/spm.py b/vdsm/storage/spm.py index b186d39..a80d03b 100644 --- a/vdsm/storage/spm.py +++ b/vdsm/storage/spm.py @@ -201,8 +201,6 @@ class SPM: blockSD.BlockStorageDomain.doUnmountMaster(master) except se.StorageDomainMasterUnmountError, e: misc.panic("unmount %s failed - %s" % (master, e)) - else: - cls.log.debug("master `%s` is not mounted, skipping", master) @classmethod def __cleanupSPMLinks(cls): @@ -1070,7 +1068,7 @@ class SPM: pool.detachSD(sdUUID, msdUUID, masterVersion) - def detachAllDomains(self, pool): + def destroyPool(self, pool): """ Detach all domains from pool before destroying pool """ @@ -1086,7 +1084,8 @@ class SPM: pool.detachSD(sdUUID=sdUUID, msdUUID=sd.BLANK_UUID, masterVersion=0) self._stop(pool.spUUID) # Forced detach 'master' domain after stopping SPM - pool.detachSD(mDom.sdUUID, sd.BLANK_UUID, 0) + pool.destroy() + return True # last master def public_attachStorageDomain(self, sdUUID, spUUID, options = None):

12 years, 11 months

2
3
0 / 0

[NEW PATCH] BZ#720425 - No longer using the return value from wait, because there is None. (via gerrit-bot)

by erez＠redhat.com

New patch submitted by Erez Sh (erez(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/684 commit b4b7c5efc10d4c0702523c69d4ac54ffba2d87b1 Author: Erez Sh <erez(a)redhat.com> Date: Mon Jul 11 18:30:48 2011 +0300 BZ#720425 - No longer using the return value from wait, because there is None. Change-Id: I1f6f22fb4e047dc24335eeef7664c33e1076eb02 diff --git a/vdsm/libvirtvm.py b/vdsm/libvirtvm.py index 0516c3b..89330c4 100644 --- a/vdsm/libvirtvm.py +++ b/vdsm/libvirtvm.py @@ -214,7 +214,8 @@ class MigrationMonitorThread(threading.Thread): def run(self): self._vm.log.debug('starting migration monitor thread') - while not self._stop.wait(self._MIGRATION_MONITOR_INTERVAL): + while not self._stop.isSet(): + self._stop.wait(self._MIGRATION_MONITOR_INTERVAL) jobType, timeElapsed, _, \ dataTotal, dataProcessed, _, \ memTotal, memProcessed, _, \

12 years, 11 months

3
3
0 / 0

Change in vdsm[master]: BZ#720425 - No longer using the return value from wait, beca...

by erez＠redhat.com

Erez Sh has posted comments on this change. Change subject: BZ#720425 - No longer using the return value from wait, because there is None. ...................................................................... Patch Set 1: Verified Ran migration tests and it worked -- To view, visit http://gerrit.usersys/684 To unsubscribe, visit http://gerrit.usersys/settings Gerrit-MessageType: comment Gerrit-Change-Id: I1f6f22fb4e047dc24335eeef7664c33e1076eb02 Gerrit-PatchSet: 1 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Erez Sh <erez(a)redhat.com> Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Erez Sh <erez(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 11 months

1
0
0 / 0

[NEW PATCH] BZ#719857 - Add reboot message to certificate reject (via gerrit-bot)

by Yotam Oron

New patch submitted by Yotam Oron (yoron(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/715 commit 50f18d38e1ab7106def66a82147bd7590e8a5e70 Author: Yotam Oron <yoron(a)redhat.com> Date: Sun Jul 17 12:07:54 2011 +0300 BZ#719857 - Add reboot message to certificate reject When the user rejects the certificate's fingerprint, the host will reboot - added a message about that. Change-Id: I6b8a60eeb56e12627e1ad3213d453a7dfb03de0d diff --git a/vdsm_reg/rhevm.py b/vdsm_reg/rhevm.py index 5ff8be6..d32c639 100755 --- a/vdsm_reg/rhevm.py +++ b/vdsm_reg/rhevm.py @@ -143,8 +143,9 @@ class Plugin(PluginBase): if deployUtil.getRhevmCert(self.rhevm_server.value(), self.rhevm_server_port.value()): path, dontCare = deployUtil.certPaths('') fp = deployUtil.generateFingerPrint(path) - approval = ButtonChoiceWindow(self.ncs.screen, \ - "Certificate Fingerprint:", fp, buttons = ['Approve', 'Reject']) + approval = ButtonChoiceWindow(self.ncs.screen, + "Certificate Fingerprint (Rejecting the finrgerprint will reboot the host):", + fp, buttons = ['Approve', 'Reject']) if 'reject' == approval: out, err, rc = deployUtil._logExec(['/sbin/reboot']) if rc is not 0:

12 years, 11 months

2
2
0 / 0

by Igor Lvovsky

New patch submitted by Igor Lvovsky (ilvovsky(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/707 commit ee4f07f9ed34f6db2aacdbc2678b4f20900ad761 Author: Igor Lvovsky <ilvovsky(a)redhat.com> Date: Tue Jul 12 16:54:31 2011 +0300 Related to BZ#713173 - Add killProc rollback to deleteImage Change-Id: I1407626ab06583f5507f3aaa8ebf2a4e4edd60f9 diff --git a/vdsm/storage/blockVolume.py b/vdsm/storage/blockVolume.py index 3da76fa..6b04047 100644 --- a/vdsm/storage/blockVolume.py +++ b/vdsm/storage/blockVolume.py @@ -261,7 +261,8 @@ class BlockVolume(volume.Volume): # wipe out the whole volume idle = config.getfloat('irs', 'idle') try: - misc.ddWatchCopy("/dev/zero", vol_path, vars.task.aborting, idle, int(size)) + misc.ddWatchCopy("/dev/zero", vol_path, vars.task.aborting, idle, int(size), + recoveryCallback=volume.baseAsyncTasksRollback) except se.ActionStopped, e: raise e except Exception, e:

12 years, 11 months

3
4
0 / 0

by Igor Lvovsky

New patch submitted by Igor Lvovsky (ilvovsky(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/709 commit bb1cba217e45a05dedcdc689611842b4b64be607 Author: Igor Lvovsky <ilvovsky(a)redhat.com> Date: Thu Jul 14 15:14:47 2011 +0300 Related to BZ#719255 - Release lock if prepare volume failed Change-Id: I0eeeda07037824bb822b8d0d1fff869959bdfefc diff --git a/vdsm/storage/hsm.py b/vdsm/storage/hsm.py index b0d5d21..ccf93d5 100644 --- a/vdsm/storage/hsm.py +++ b/vdsm/storage/hsm.py @@ -1671,8 +1671,8 @@ class HSM: vol.refreshVolume() vol.prepare(rw=rw) except: - imgResource.autoRelease = True self.log.error("Prepare volume %s in domain %s failed", volUUID, sdUUID, exc_info=True) + rmanager.releaseResource(imageResourcesNamespace, imgUUID) raise

12 years, 11 months

3
8
0 / 0

[NEW PATCH] BZ#720670 vdsm-reg: rotate correct log file (via gerrit-bot)

by Dan Kenigsberg

New patch submitted by Dan Kenigsberg (danken(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/694 commit b033abf7a59837d2e15d191049b0f6a23dc716cc Author: Robert Munilla <rmunilla(a)redhat.com> Date: Tue Jul 12 16:58:03 2011 +0300 BZ#720670 vdsm-reg: rotate correct log file Change-Id: Ic950727ba692317be5172053e5e62d01805b7ecd diff --git a/vdsm_reg/vdsm-reg-logrotate.conf b/vdsm_reg/vdsm-reg-logrotate.conf index 45572d7..3b590dd 100644 --- a/vdsm_reg/vdsm-reg-logrotate.conf +++ b/vdsm_reg/vdsm-reg-logrotate.conf @@ -4,7 +4,7 @@ maxage 30 } -/var/log/vdsm-reg/vds-reg.log { +/var/log/vdsm-reg/vdsm-reg.log { rotate 3 missingok size=10M

12 years, 11 months

2
3
0 / 0

[NEW PATCH] BZ#717847 - Postpone VM's periodic tasks until all volumes will prepared (via gerrit-bot)

by Igor Lvovsky

New patch submitted by Igor Lvovsky (ilvovsky(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/697 commit 7c44b0af0303d7a4876091cc5bfc9783dbcb5ecb Author: Igor Lvovsky <ilvovsky(a)redhat.com> Date: Wed Jul 13 15:14:19 2011 +0300 BZ#717847 - Postpone VM's periodic tasks until all volumes will prepared Change-Id: I9424087c3c9ef06a4a88921dc8ca14e890b7a44f diff --git a/vdsm/libvirtvm.py b/vdsm/libvirtvm.py index 40567a5..79da4b2 100644 --- a/vdsm/libvirtvm.py +++ b/vdsm/libvirtvm.py @@ -55,6 +55,10 @@ class VmStatsThread(utils.AdvancedStatsThread): if self._vm._incomingMigrationPending(): return + if not self._vm._volumesPrepared: + # Avoid queries from storage during recovery process + return + for vmDrive in self._vm._drives: if vmDrive.blockDev and vmDrive.format == 'cow': capacity, alloc, physical = \ @@ -66,7 +70,7 @@ class VmStatsThread(utils.AdvancedStatsThread): self._vm._onHighWrite(vmDrive.name, alloc) def _updateVolumes(self): - if not self._vm.cif.irs.getConnectedStoragePoolsList()['poollist']: + if not self._vm._volumesPrepared: # Avoid queries from storage during recovery process return @@ -82,6 +86,10 @@ class VmStatsThread(utils.AdvancedStatsThread): return cpuTime / 1000**3 def _sampleDisk(self): + if not self._vm._volumesPrepared: + # Avoid queries from storage during recovery process + return + diskSamples = {} for vmDrive in self._vm._drives: diskSamples[vmDrive.name] = self._vm._dom.blockStats(vmDrive.name) diff --git a/vdsm/vm.py b/vdsm/vm.py index ecc8e70..5b30b1e 100644 --- a/vdsm/vm.py +++ b/vdsm/vm.py @@ -300,6 +300,7 @@ class Vm(object): self.conf.pop('elapsedTimeOffset', 0)) self._cdromPreparedPath = '' self._floppyPreparedPath = '' + self._volumesPrepared = False self._pathsPreparedEvent = threading.Event() self.saveState() @@ -418,6 +419,8 @@ class Vm(object): def preparePaths(self): for drive in self.conf.get('drives', []): drive['path'] = self._prepareVolumePath(drive) + # Now we got all needed locks + self._volumesPrepared = True try: self._cdromPreparedPath = self._prepareVolumePath(

12 years, 11 months

3
4
0 / 0

[NEW PATCH] BZ#720919 spmprotect.sh: remove log noise (via gerrit-bot)

by Dan Kenigsberg

New patch submitted by Dan Kenigsberg (danken(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/696 commit cfcabd3b7bba2d6513b65216017828f326f2f660 Author: Dan Kenigsberg <danken(a)redhat.com> Date: Wed Jul 13 14:23:45 2011 +0300 BZ#720919 spmprotect.sh: remove log noise check_renew may have been called after fence or release have removed RENEWDIR. We should not print nonsense to the log in that case. Change-Id: I7ce575b863fcbdf6b0bf35747a23882a53ae1a39 diff --git a/vdsm/storage/protect/spmprotect.sh b/vdsm/storage/protect/spmprotect.sh index 3548b17..f476446 100755 --- a/vdsm/storage/protect/spmprotect.sh +++ b/vdsm/storage/protect/spmprotect.sh @@ -135,7 +135,7 @@ function check_renew() { res=0 fi fi - if pushd "$RENEWDIR" > /dev/null; then + if pushd "$RENEWDIR" > /dev/null 2>&1 ; then rm -f $list popd > /dev/null fi

12 years, 11 months

2
4
0 / 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

vdsm-patches July 2011