Red Hat Bugzilla – Bug 822393
[vdsm] getVGInfo on specific VG fails due to access problem with different VG
Last modified: 2014-01-12 19:52:05 EST
Description of problem:
getVGInfo command on newly created storage fails due to access problem with one
of the other VGs in the system.
- create new storage domain using RHEVM
* createVG command - returns successfully
* createStorageDomain command - returns successfully
* getVGInfo on vgUUID - fails;
AttributeError: Failed reload: 703256d8-0c4a-4138-a150-e4db0ad87db9
when I issue vgs command, I see VG, also, getStorageDomainInfo command returns.
problem solved just after I restarted vdsm, which points on cache issue.
also, i'm curious why getVGInfo, which requests specific device, fails, in case
different and unrelated vg is down.
attached vdsm log.
we have a patch on downstream\upstream:
git show fb7012c7
Author: Eduardo Warszawski <email@example.com>
Date: Tue Sep 13 10:39:31 2011 +0300
BZ#736301 - Don't fail lvm.getVGbyUUID() if unreadable VGs found.
Reviewed-by: Dan Kenigsberg <firstname.lastname@example.org>
Tested-by: Dan Kenigsberg <email@example.com>
diff --git a/vdsm/storage/lvm.py b/vdsm/storage/lvm.py
index 0276f9e..9d734f7 100644
@@ -833,12 +833,18 @@ def getVGs(vgNames):
return _lvminfo.getAllVgs() #returns list
+# TODO: lvm VG UUID should not be exposed.
+# Remove this function when hsm.public_createVG is removed.
# cycle through all the VGs until the one with the given UUID found
for vg in getAllVGs():
- if vg.uuid == vgUUID:
- return vg
+ if vg.uuid == vgUUID:
+ return vg
+ except AttributeError, e:
+ # An unreloadable VG found but may be we are not looking for it.
+ log.debug("%s" % e.message, exc_info=True)
# If not cry loudly
raise se.VolumeGroupDoesNotExist("vg_uuid: %s" % vgUUID)
This request was not resolved in time for the current release.
Red Hat invites you to ask your support representative to
propose this request, if still desired, for consideration in
the next release of Red Hat Enterprise Linux.
(In reply to comment #0)
> attached vdsm log.
Actually, no. Since we have the patch for upstream and form rhev-3.1, I am guessing you are refering to rhev-3.0.
I do not mind making this bug track this issue for the next 3.0.z build. Setting the Zstream keyword, as the 3.1 bug 736301 is already fixed.