September 2011 - vdsm-patches - Fedora Mailing-Lists

Change in vdsm[master]: Don't use a list to calculate VG size

by Dan Kenigsberg

Dan Kenigsberg has posted comments on this change. Change subject: Don't use a list to calculate VG size ...................................................................... Patch Set 2: Verified; Looks good to me, approved -- To view, visit http://gerrit.usersys.redhat.com/967 To unsubscribe, visit http://gerrit.usersys.redhat.com/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ief2a3e69f94c99b11c9027901fc5a34577c3452f Gerrit-PatchSet: 2 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Igor Lvovsky <ilvovsky(a)redhat.com> Gerrit-Reviewer: Yotam Oron <yoron(a)redhat.com>

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#732980 Devices block size must be 512

by Dan Kenigsberg

Dan Kenigsberg has posted comments on this change. Change subject: BZ#732980 Devices block size must be 512 ...................................................................... Patch Set 9: Verified; Looks good to me, approved -- Gerrit-MessageType: comment Gerrit-Change-Id: If303fac85a29bed7b989b0b6302fc4c83075f500 Gerrit-PatchSet: 9 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Ayal Baron Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#740887 Tune cache dirty ratio

by Dan Kenigsberg

Dan Kenigsberg has posted comments on this change. Change subject: BZ#740887 Tune cache dirty ratio ...................................................................... Patch Set 1: I would prefer that you didn't submit this (2 inline comments) .................................................... Commit Message Line 3: AuthorDate: 2011-09-26 15:28:36 +0000 Line 4: Commit: Federico Simoncelli <fsimonce(a)redhat.com> Line 5: CommitDate: 2011-09-26 15:32:34 +0000 Line 6: Line 7: BZ#740887 Tune cache dirty ratio Could you explain why we need to do it? bugzilla does not give to much, either. Line 8: Line 9: Change-Id: Ibf5c8e4c0637c60092b89fba103b96b37bdafaa0 Line 10 Line 11 .................................................... File vdsm/clientIF.py Line 143: self.irs.prepareForShutdown() Line 144: raise Line 145: Line 146: def _tuneDirtyRatio(self): Line 147: storage.misc.setDirtyRatio(config.get('vars', 'dirty_ratio')) storage.misc? why? and if so, it should be [irs] section, probably. Line 148: storage.misc.setDirtyBackgroundRatio( Line 149: config.get('vars', 'dirty_background_ratio')) Line 150: Line 151: def _getServerIP(self): -- To view, visit http://10.35.18.144/970 To unsubscribe, visit http://10.35.18.144/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ibf5c8e4c0637c60092b89fba103b96b37bdafaa0 Gerrit-PatchSet: 1 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com>

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#732980 Devices block size must be 512

by Federico Simoncelli

Federico Simoncelli has posted comments on this change. Change subject: BZ#732980 Devices block size must be 512 ...................................................................... Patch Set 8: (1 inline comment) .................................................... File vdsm/storage/multipath.py Line 254: except Exception: Line 255: log.warn("Problem getting fwrev from device `%s`", slave, exc_info=True) Line 256: Line 257: if not devInfo["logicalblocksize"] or not devInfo["physicalblocksize"]: Line 258: try: Who is not reporting the block size? The kernel? All block devices must report a block size, if they don't we fail. Line 259: logBlkSize, phyBlkSize = getDeviceBlockSizes(slave) Line 260: devInfo["logicalblocksize"] = str(logBlkSize) Line 261: devInfo["physicalblocksize"] = str(phyBlkSize) Line 262: except Exception: -- To view, visit http://10.35.18.144/855 To unsubscribe, visit http://10.35.18.144/settings Gerrit-MessageType: comment Gerrit-Change-Id: If303fac85a29bed7b989b0b6302fc4c83075f500 Gerrit-PatchSet: 8 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Ayal Baron Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#732980 Devices block size must be 512

by Dan Kenigsberg

Dan Kenigsberg has posted comments on this change. Change subject: BZ#732980 Devices block size must be 512 ...................................................................... Patch Set 8: Verified; Looks good to me, approved -- Gerrit-MessageType: comment Gerrit-Change-Id: If303fac85a29bed7b989b0b6302fc4c83075f500 Gerrit-PatchSet: 8 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Ayal Baron Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 7 months

1
0
0 / 0

[MERGED PATCH] BZ#732980 Check block size on VG commands (via gerrit-bot)

by Federico Simoncelli

Patch merged into repository You can review this change at: http://gerrit.usersys.redhat.com/920

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#732980 Check block size on VG commands

by Dan Kenigsberg

Dan Kenigsberg has submitted this change and it was merged. Change subject: BZ#732980 Check block size on VG commands ...................................................................... BZ#732980 Check block size on VG commands Change-Id: Ie106491328f12f650b6bb234a92238d6c0aa70c8 --- M vdsm/storage/blockSD.py M vdsm/storage/lvm.py M vdsm/storage/storage_exception.py 3 files changed, 31 insertions(+), 0 deletions(-) Approvals: Dan Kenigsberg: Verified; Looks good to me, approved Ayal Baron: Looks good to me, but someone else must approve Saggi Mizrahi: Looks good to me, but someone else must approve -- Gerrit-MessageType: merged Gerrit-Change-Id: Ie106491328f12f650b6bb234a92238d6c0aa70c8 Gerrit-PatchSet: 4 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Ayal Baron Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 7 months

1
0
0 / 0

Change in vdsm[master]: BZ#732980 Check block size on VG commands

by Dan Kenigsberg

Dan Kenigsberg has posted comments on this change. Change subject: BZ#732980 Check block size on VG commands ...................................................................... Patch Set 4: Verified; Looks good to me, approved -- Gerrit-MessageType: comment Gerrit-Change-Id: Ie106491328f12f650b6bb234a92238d6c0aa70c8 Gerrit-PatchSet: 4 Gerrit-Project: vdsm Gerrit-Branch: master Gerrit-Owner: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Ayal Baron Gerrit-Reviewer: Dan Kenigsberg <danken(a)redhat.com> Gerrit-Reviewer: Federico Simoncelli <fsimonce(a)redhat.com> Gerrit-Reviewer: Saggi Mizrahi <smizrahi(a)redhat.com>

12 years, 7 months

1
0
0 / 0

[MERGED PATCH] BZ#732980 MD tag must be in blocks unit (via gerrit-bot)

by Federico Simoncelli

Patch merged into repository You can review this change at: http://gerrit.usersys.redhat.com/891

12 years, 7 months

1
0
0 / 0

[NEW PATCH] BZ#732980 MD tag must be in blocks unit (via gerrit-bot)

by Federico Simoncelli

New patch submitted by Federico Simoncelli (fsimonce(a)redhat.com) You can review this change at: http://gerrit.usersys.redhat.com/891 commit 1a3aa6f6cbf9913ff3ad8e8e481b519e54853bfe Author: Federico Simoncelli <fsimonce(a)redhat.com> Date: Fri Sep 2 16:36:33 2011 +0000 BZ#732980 MD tag must be in blocks unit This patch fixes the unit used for the MD tag (blocks instead of bytes) and completes the support for the new MS tag which holds the size (in blocks) of the metadata. Change-Id: Id268735f60a289f9e8f12a7d7bb3180bf3398c54 diff --git a/vdsm/storage/blockSD.py b/vdsm/storage/blockSD.py index ab3241f..c5aa4a9 100644 --- a/vdsm/storage/blockSD.py +++ b/vdsm/storage/blockSD.py @@ -520,11 +520,11 @@ class BlockStorageDomain(sd.StorageDomain): finally: self._extendlock.release() - def mapMetaOffset(self, vol_name): + def mapMetaOffset(self, vol_name, slotSize): if self.getVersion() in VERS_METADATA_LV: return self.getVolumeMetadataOffsetFromPvMapping(vol_name) else: - return self.getFreeMetadataSlot(blockVolume.VOLUME_METASIZE) + return self.getFreeMetadataSlot(slotSize) def _getOccupiedMetadataSlots(self): stripPrefix = lambda s, pfx : s[len(pfx):] @@ -540,6 +540,9 @@ class BlockStorageDomain(sd.StorageDomain): if tag.startswith(blockVolume.TAG_PREFIX_MD): offset = int(stripPrefix(tag, blockVolume.TAG_PREFIX_MD)) + if tag.startswith(blockVolume.TAG_PREFIX_MS): + size = int(stripPrefix(tag, blockVolume.TAG_PREFIX_MS)) + if offset is not None and size is not None: # I've found everything I need break @@ -549,7 +552,8 @@ class BlockStorageDomain(sd.StorageDomain): continue if size is None: - size = blockVolume.VOLUME_METASIZE + self.log.warn("Could not find size for lv %s/%s", self.sdUUID, lv.name) + continue occupiedSlots.append((offset, size)) @@ -558,14 +562,21 @@ class BlockStorageDomain(sd.StorageDomain): def getFreeMetadataSlot(self, slotSize): occupiedSlots = self._getOccupiedMetadataSlots() + logBlkSize, phyBlkSize = lvm.getVGBlockSizes(self.sdUUID) + + # It might look weird skipping the sd metadata when it has been moved + # to tags. But this is here because domain metadata and volume metadata + # look the same. The domain might get confused and think it has lv + # metadata if it finds something is written in that area. + # At the moment this check is trivial but it will be needed in the + # future when we'll support multiple block sizes. + if logBlkSize >= SD_METADATA_SIZE: + freeSlot = 1 + else: + freeSlot, freeSlotRem = divmod(SD_METADATA_SIZE, logBlkSize) + if freeSlotRem > 0: + freeSlot += 1 - # It might look weird skipping the sd metadata - # when it has been moved to tags. But this is - # here because domain metadata and volume - # metadata look the same. The domain might get - # confused and think it has lv metadata if it - # finds something is written in that area. - freeSlot = SD_METADATA_SIZE for offset, size in occupiedSlots: if offset - freeSlot > slotSize: break diff --git a/vdsm/storage/blockVolume.py b/vdsm/storage/blockVolume.py index 25ec92c..a83b1f7 100644 --- a/vdsm/storage/blockVolume.py +++ b/vdsm/storage/blockVolume.py @@ -38,15 +38,18 @@ from resourceFactories import LVM_ACTIVATION_NAMESPACE import fileUtils TAG_PREFIX_MD = "MD_" +TAG_PREFIX_MS = "MS_" TAG_PREFIX_IMAGE = "IU_" TAG_PREFIX_PARENT = "PU_" VOLUME_TAGS = [TAG_PREFIX_PARENT, TAG_PREFIX_IMAGE, - TAG_PREFIX_MD] + TAG_PREFIX_MD, + TAG_PREFIX_MS] # volume meta data block size VOLUME_METASIZE = 512 +VOLUME_METASLOT = 1 rmanager = rm.ResourceManager.getInstance() @@ -215,8 +218,9 @@ class BlockVolume(volume.Volume): try: with cls._tagCreateLock: - offs = mysd.mapMetaOffset(volUUID) + offs = mysd.mapMetaOffset(volUUID, VOLUME_METASLOT) lvm.addLVTags(sdUUID, volUUID, ("%s%s" % (TAG_PREFIX_MD, offs), + "%s%s" % (TAG_PREFIX_MS, VOLUME_METASLOT), "%s%s" % (TAG_PREFIX_PARENT, srcVolUUID,), "%s%s" % (TAG_PREFIX_IMAGE, imgUUID,))) diff --git a/vdsm/storage/lvm.py b/vdsm/storage/lvm.py index 921cf3b..f1b25a3 100644 --- a/vdsm/storage/lvm.py +++ b/vdsm/storage/lvm.py @@ -934,6 +934,14 @@ def invalidateVG(vgName): _lvminfo._invalidatevgs(vgName) _lvminfo._invalidatelvs(vgName) +def getVGBlockSizes(vgUUID): + pvs = [pv.name for pv in _lvminfo.getAllPvs() if pv.vg_name == vgUUID] + if not pvs: + return None + # Returning the block size of the first pv is correct since we don't allow + # devices with different block size to be on the same VG. + dmDevice = devicemapper.getDmId(os.path.basename(pvs[0])) + return multipath.getDeviceBlockSizes(dmDevice) # # Public Logical volume interface

12 years, 7 months

3
8
0 / 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

vdsm-patches September 2011