Bug 1075263 (BrassowRHEL6Bugs)

Summary: Brassow RHEL6 tracker
Product: Red Hat Enterprise Linux 6 Reporter: Jonathan Earl Brassow <jbrassow>
Component: lvm2Assignee: Jonathan Earl Brassow <jbrassow>
Status: CLOSED NOTABUG QA Contact: Jonathan Earl Brassow <jbrassow>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 6.5CC: jbrassow
Target Milestone: rcKeywords: Tracking
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2015-09-29 21:59:31 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 669111, 692186, 811348, 829920, 834050, 853247, 875930, 877221, 894334, 901697, 905063, 960396, 995157, 1014758, 1021955, 1024474, 1069929, 1085904, 1095834, 1095843, 1102564    
Bug Blocks: 958235    

Description Jonathan Earl Brassow 2014-03-11 21:50:42 UTC
Just using this bug to group and prioritize other bugs.

Comment 2 Jonathan Earl Brassow 2014-03-11 22:46:20 UTC
RHEL6 bugs in priority order:

- Bug 1069929 - DM RAID: Device recovery skipped by cycling the logical volume

- Bug 901697 - 'Parse error at byte 77 (line 5): unexpected token' during multiple raid volume creation

- Bug 1031204 - raid volumes experiencing multiple device failures can see repair failures

- Bug 811348 - raid down conversion gives confusing message when specifying too few images to remove

- Bug 829920 - mirror and raid creation should be able to use -l100%FREE

- Bug 1021955 - Copy/Sync percentage drops when maxrecoveryrate is changed on a RAID LV

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 875930 - vgsplit can not split two device mirrors where the log and a leg share the same device

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

Comment 3 Jonathan Earl Brassow 2014-04-08 21:20:39 UTC
Bug 1069929:  Kernel patch identified.  Jes to post with other MD updates.
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 811348 :  POST.  Message improved.
Bug 1031204:  Unable to reproduce - even with old 6.5 RPMs
Bug 1021955:  Unable to reproduce - even with old 6.5 RPMs

New priority list
=================
- Bug 1031204 - raid volumes experiencing multiple device failures can see repair failures

- Bug 829920 - mirror and raid creation should be able to use -l100%FREE

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 875930 - vgsplit can not split two device mirrors where the log and a leg share the same device

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

- Bug 1069929 - Waiting for kernel patch submission
- Bug 1031204:  Waiting for QA to reproduce.
- Bug 1021955:  Waiting for QA to reproduce.

Comment 4 Jonathan Earl Brassow 2014-04-22 19:44:11 UTC
Bug 829920 :  POST.  Divided into two bugs - one for RAID (bug 829920) and one for mirror (bug 1085904).  The one for mirror is deferred to 6.7.

Comment 5 Jonathan Earl Brassow 2014-04-22 20:06:46 UTC
Bug 1069929:  Kernel patch identified.  Jes to post with other MD updates.
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 1031204:  Unable to reproduce - even with old 6.5 RPMs

New priority list
=================
- Bug 1085553 - When lvmetad is used, LVs do not properly report as 'p'artial
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 875930 - vgsplit can not split two device mirrors where the log and a leg share the same device

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

- Bug 1069929 - Waiting for kernel patch submission
- Bug 1031204:  Waiting for QA to reproduce.

Comment 6 Jonathan Earl Brassow 2014-05-07 04:47:46 UTC
Bug 1069929:  Kernel patch identified.  Jes to post with other MD updates.
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 875930 :  POST.  Also revealed RAID 4/5/6 vgsplit bug that is now fixed.
Bug 1031204:  Unable to reproduce - even with old 6.5 RPMs
Bug 1085553:  Reassigned to mornfall.  (Implies bug 1089170 and bug 1089369.)

New priority list
=================
- Bug 669111 - [RFE] pvmove complete PVs atomically

- Bug 960396 - Pvmove on clustered VG should check the modules and the service on all the nodes

- Bug 1014758 - LVM RAID + thin do not work in a cluster (even when activated exclusively) [Need to reassign to zkabelac for activation changes?]

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 995157 - [RHEL 6.4] LVM allows setting of clustered attribute on VG even though clvmd or cluster configured or running.

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

- Bug 1069929 - Waiting for kernel patch submission
- Bug 1031204:  Waiting for QA to reproduce.
- Bug 1085553 - Will keep watching - 'lvs' not properly reporting RAID 'p'
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)

Comment 7 Jonathan Earl Brassow 2014-05-30 18:51:25 UTC
Bug 1069929:  MODIFIED.  Patch is in kernel-2.6.32-466.el6.
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 875930 :  POST.  Also revealed RAID 4/5/6 vgsplit bug that is now fixed.
Bug 960396 :  POST.  Suitable changes made to satisfy request.
Bug 1031204:  CLOSED WORKSFORME.  Unable to reproduce.
Bug 1085553:  Reassigned to mornfall.  (Implies bug 1089170 and bug 1089369.)

New priority list
=================
- Bug 1102564 - lvchange --syncaction is not detecting corruptions

- Bug 669111 - [RFE] pvmove complete PVs atomically.  [I've got a POC patch that works for some simple cases, but not all cases are yet covered - like when allocations need splitting or there are many LV to handle.]

- Bug 1014758 - LVM RAID + thin do not work in a cluster (even when activated exclusively) [Need to reassign to zkabelac for activation changes?]

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 995157 - [RHEL 6.4] LVM allows setting of clustered attribute on VG even though clvmd or cluster configured or running.

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

- Bug 1095834 - LVM RAID/Mirror: splitmirror should automatically flush the file system

- Bug 1095843 - LVM cache: Not all sub-LVs renamed when caching a RAID LV

- Bug 692186 - HA LVM mirror repair can print worrisome messages on device failure + service relocation

- Bug 905063 - Multiple issues with lvm2 and 2-disk mirrored configurations

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

- Bug 1085553 - Will keep watching - 'lvs' not properly reporting RAID 'p'
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)

Comment 8 Jonathan Earl Brassow 2014-06-03 22:17:38 UTC
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 875930 :  POST.  Also revealed RAID 4/5/6 vgsplit bug that is now fixed.
Bug 960396 :  POST.  Suitable changes made to satisfy request.
Bug 1085553:  Reassigned to mornfall.  (Implies bug 1089170 and bug 1089369.)
Bug 1014758:  Reassigned to zkabelac.  (Activation issues.)
Bug 1102564:  CLOSED NOTABUG.  Updated testsuite.

New priority list
=================
- Bug 669111 - [RFE] pvmove complete PVs atomically.  [I've got a POC patch that works for some simple cases, but not all cases are yet covered - like when allocations need splitting or there are many LV to handle.]

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 995157 - [RHEL 6.4] LVM allows setting of clustered attribute on VG even though clvmd or cluster configured or running.

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss

- Bug 1095834 - LVM RAID/Mirror: splitmirror should automatically flush the file system [have patch - needs review]

- Bug 1095843 - LVM cache: Not all sub-LVs renamed when caching a RAID LV [have patch - needs review]

- Bug 905063 - Multiple issues with lvm2 and 2-disk mirrored configurations

- Bug 692186 - HA LVM mirror repair can print worrisome messages on device failure + service relocation [have patch - needs review]

- Bug 877221 - lvconvert --repair won't reuse physical volumes

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes

- Bug 1014758 - Will keep watching - RAID + thin cluster activation issues
- Bug 1085553 - Will keep watching - 'lvs' not properly reporting RAID 'p'
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)

Comment 9 Jonathan Earl Brassow 2014-06-05 22:20:57 UTC
Bug 905063 is a 6.6UserspacePrio bug, but I am looking to close WONTFIX.

Comment 10 Jonathan Earl Brassow 2014-06-06 03:52:42 UTC
Bug 995157 now in POST.  Note, there are two upstream commits.

Comment 11 Jonathan Earl Brassow 2014-06-06 05:46:58 UTC
Patches for bug 1095843 and bug 692186 sent to lvm-devel for feedback.

Comment 12 Jonathan Earl Brassow 2014-06-16 20:51:58 UTC
LVM code changes causing regressions in cache:
bug 1107345 - LVM cache: Unable to remove cache-pool volume
bug 1104381 - RAID cache: Removing cache logical volume prints stray error message
bug 1110026 - LVM cache: Unable to lvcreate cache LV from existing origin LV

Comment 13 Jonathan Earl Brassow 2014-06-18 04:42:15 UTC
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 875930 :  POST.  Also revealed RAID 4/5/6 vgsplit bug that is now fixed.
Bug 960396 :  POST.  Suitable changes made to satisfy request.
Bug 692186 :  POST.  Cleanly exit poll daemoon - no scary msg.
Bug 1095843:  POST.  Proper rename of RAID sub-lvs when converting to cache.
Bug 995157 :  POST.  Query user for cluster flag change if cluster not setup
Bug 669111 :  POST.  Atomic pvmove.
Bug 905063 :  ASSIGNED.  Letting ruzou close for me after testing. [3ACK]
Bug 1102564:  CLOSED NOTABUG.  Updated testsuite.

New priority list
=================
- Bug 1102564 - lvchange --syncaction is not detecting corruptions [needs QA ack]

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed [3ACK]

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss [3ACK]

- Bug 1095834 - LVM RAID/Mirror: splitmirror should automatically flush the file system [have patch - needs clean-up before review]

- Bug 877221 - lvconvert --repair won't reuse physical volumes [PM PRIORITY LIST][3ACK]

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes [3ACK]

Others' bugs I'm watching
=========================
- Bug 1107345:  Cache regression.
- Bug 1104381:  Cache regression.
- Bug 1110026:  Cache regression.
- Bug 1014758 - RAID + thin cluster activation issues
- Bug 1085553 - 'lvs' not properly reporting RAID 'p'
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)

Comment 14 Jonathan Earl Brassow 2014-06-26 03:57:33 UTC
Bug 901697 :  POST.  Retest shows message has changed and is sufficient.
Bug 829920 :  POST.
Bug 811348 :  POST.  Message improved.
Bug 875930 :  POST.  Also revealed RAID 4/5/6 vgsplit bug that is now fixed.
Bug 960396 :  POST.  Suitable changes made to satisfy request.
Bug 692186 :  POST.  Cleanly exit poll daemoon - no scary msg.
Bug 1095843:  POST.  Proper rename of RAID sub-lvs when converting to cache.
Bug 995157 :  POST.  Query user for cluster flag change if cluster not setup
Bug 669111 :  POST.  Atomic pvmove.
Bug 877221 :  POST.  'lvconvert --repair' reuse PVs.
Bug 905063 :  ASSIGNED.  Letting ruzou close for me after testing. [3ACK]

New priority list
=================
- Bug 669111 - Atomic pvmove [still needs persistent log][have patch]

- Bug 905063 - Multiple issues with lvm2 and 2-disk mirrored configurations [RPL]

- Bug 853247 - lvm resource fails to start/stop if owning node rejoins before token loss [3ACK][HA TEAM PRIORITY]

- Bug 1102564 - lvchange --syncaction is not detecting corruptions [needs QA ack]

- Bug 894334 - mirror leg not restored after failure: lvm[5786]: Secondary mirror device 253:4 sync failed [3ACK]

- Bug 1024474 - vgchange –a y –partial” command hangs indefinitely when some of the physical volumes are not accessible

- Bug 1095834 - LVM RAID/Mirror: splitmirror should automatically flush the file system [have patch - needs clean-up before review]

- Bug 834050 - Unable to create striped raid on VGs with 1k extent sizes [3ACK]

Others' bugs I'm watching
=========================
- Bug 1107345:  Cache regression.
- Bug 1104381:  Cache regression.
- Bug 1110026:  Cache regression.
- Bug 1014758 - RAID + thin cluster activation issues
- Bug 1085553 - 'lvs' not properly reporting RAID 'p'
-- bug 1089170 - A duplicate (resulting mirror issues)
-- bug 1089369 - A duplicate (resulting RAID issues)