Red Hat Bugzilla – Bug 481384
RFE: allow the ability to stack pvs on top of existing mirrors
Last modified: 2011-02-09 09:48:49 EST
Description of problem:
This is a pretty low priority bug, it would be nice to be able to stack pvs on top of mirrors like you can linears. You can not however because the volume header get written to both places and that throws lvm for a loop. You'll end up with "Found duplicate PV" messages and it wont work.
This is possible if you use filters however. You can filter out mimage* and mlog*.
[root@hayes-03 ~]# pvscan
Found duplicate PV mK9mPc21WClvcWPtIcGQPQqgHHTzAf6U: using /dev/mapper/vg2-mirror2_mimage_1 not /dev/mapper/vg2-mirror2_mimage_0
Found duplicate PV mK9mPc21WClvcWPtIcGQPQqgHHTzAf6U: using /dev/vg2/mirror2 not /dev/mapper/vg2-mirror2_mimage_1
Found duplicate PV zQOZSXuN51TwwrCBU6miKAed3EMod4U0: using /dev/mapper/vg1-mirror1_mimage_1 not /dev/mapper/vg1-mirror1_mimage_0
Found duplicate PV zQOZSXuN51TwwrCBU6miKAed3EMod4U0: using /dev/vg1/mirror1 not /dev/mapper/vg1-mirror1_mimage_1
PV /dev/etherd/e1.1p3 VG vg2 lvm2 [1.77 TB / 1.77 TB free]
PV /dev/etherd/e1.1p4 VG vg2 lvm2 [1.77 TB / 1.77 TB free]
PV /dev/etherd/e1.1p1 VG vg1 lvm2 [1.77 TB / 1.77 TB free]
PV /dev/etherd/e1.1p2 VG vg1 lvm2 [1.77 TB / 1.77 TB free]
PV /dev/vg2/mirror2 VG stacked lvm2 [1020.00 MB / 1020.00 MB free]
PV /dev/vg1/mirror1 VG stacked lvm2 [1020.00 MB / 1008.00 MB free]
PV /dev/etherd/e1.1p5 VG stacked lvm2 [1.77 TB / 1.77 TB free]
PV /dev/sda2 VG VolGroup00 lvm2 [74.38 GB / 0 free]
Total: 8 [8.94 TB] / in use: 8 [8.94 TB] / in no VG: 0 [0 ]
Version-Release number of selected component (if applicable):
lvm2-2.02.40-6.el5 BUILT: Fri Oct 24 07:37:33 CDT 2008
lvm2-cluster-2.02.40-7.el5 BUILT: Wed Nov 26 07:19:19 CST 2008
device-mapper-1.02.28-2.el5 BUILT: Fri Sep 19 02:50:32 CDT 2008
Apply filter in the config file and it will work. This is really not a bug, but incorrect configuration.
Found duplicate PV mK9mPc21WClvcWPtIcGQPQqgHHTzAf6U: using
/dev/mapper/vg2-mirror2_mimage_1 not /dev/mapper/vg2-mirror2_mimage_0
Well the code did choose the correct devices to use - just the messages are ugly.
(Did the recent changes concerning not scanning hidden devices in 2.02.73 fix this already?)
(If so, this one is POST.)
Mornfall, I assume Alasdair is referring to your "Never scan internal LVM devices." code.
It should be fixed indeed. Since the fix went into .73, I would believe that it is even fixed in current 5.6 build.
Already fixed some time ago: closing.