1 |
Hi Paul, |
2 |
|
3 |
as an LVM user without deeper knowledge, stumbled upon your question without |
4 |
any reply yet, and my fear was to run into the same when updating lvm2. |
5 |
|
6 |
On 5/17/19 2:26 PM, Paul Colquhoun wrote: |
7 |
> Recently I found that new kernels were not booting for me, because they could |
8 |
> not assemble the LVM partition that I use for the root filesystem. |
9 |
> |
10 |
> Booting back to my old kernel still worked. |
11 |
> |
12 |
> I have tracked this back to the lvm2 version. |
13 |
> |
14 |
> After booting with the old kernel, I ran lvm and tried the 'fullreport' |
15 |
> command. |
16 |
> |
17 |
> sys-fs/lvm2-2.02.184-r3 gives an error: |
18 |
> |
19 |
> lvm> fullreport |
20 |
> LV root invalid: visible raid meta LV for raid1 segment |
21 |
> LV root invalid: visible raid meta LV for raid1 segment |
22 |
> Internal error: LV segments corrupted in root. |
23 |
|
24 |
Searching the web with parts of this error messages leads me to this commit: |
25 |
https://github.com/lvmteam/lvm2/commit/dd5716ddf258c4a44819fa90d3356833ccf767b4 |
26 |
|
27 |
While I have no idea about "visible SubLVs", maybe that commit message can |
28 |
tell something to you? |
29 |
|
30 |
> |
31 |
> After backing out to an earlier version, sys-fs/lvm2-2.02.183 |
32 |
> the 'fullreport' actually gives a report. |
33 |
> |
34 |
> I'm assuming the only reason the old kernel boots is that it has the older lvm |
35 |
> in the initramfs, and once assembled the handover to the live system still |
36 |
> works. |
37 |
> |
38 |
> I can't find anything online that looks like the same thing to me, so I was |
39 |
> wondering if anyone here had encountered a similar problem? |
40 |
> |
41 |
> The next step is to try and find how to update the on-disk lvm meta data so the |
42 |
> later versions understand it, hopefully without having to rebuild my system |
43 |
> from scratch. |
44 |
|
45 |
As far as I understand, this doesn't seem like a metadata format _change_, but |
46 |
a rare metadata consistency problem that goes unnoticed by the older version. |
47 |
|
48 |
HTH, |
49 |
/haubi/ |