[dm-devel] multipath-0.5.0 still provides broken udev rules

[Peter Volkov]

On Mon, Apr 25, 2016 at 1:58 PM, Zdenek Kabelac <zkabelac at redhat.com> wrote:

> On 25.4.2016 12:52, Peter Volkov wrote:
>>
>> There is a problem: udev does not create partitions for multipath devices
>> in
>> case I use kpartx.rules provided with multipath sources. I found that udev
>> always go to kpartx_end after following line of rules:
>> ENV{DM_TABLE_STATE}!="LIVE", GOTO="kpartx_end"
>>
>
> Just curious - what would you want to do with inactive table???
>

Actually don't understand what's this table and what's wrong with that.
There is no traces of DM_TABLE_STATE variable in lvm2 sources. Yet there is
such variable in udev rules file that comes with multipath sources:
http://git.opensvc.com/gitweb.cgi?p=multipath-tools/.git;a=blob;f=kpartx/kpartx.rules;h=022361f907e873ac16fc75459b88af34b27576e5;hb=HEAD


> You cannot read/scan such device ?


The problem I have - multipath with udev's help creates /dev/mapper/mpatha
block device but does not create devices for partitions, e.g.
/dev/mapper/mpatha1. I found that this happens due to udev rule mentioned
above. After execution rule on line 10 (*ENV{DM_TABLE_STATE}!="LIVE",
GOTO="kpartx_end"*) udev goes directly to kpartx_end and rule on line 45
(..."/sbin/kpartx -u -p -part /dev/$name") is never executed.


> There is 8 (!!!) years old thread that discusses this problem:
>> https://www.redhat.com/archives/dm-devel/2007-July/msg00095.html
>>
>> So what is the current state? Should we fix udev rules for multipath or
>> should
>> we create better patch for dmsetup.c? Meanwhile I've copied
>> multipath.rules
>> from redhat and partitions are created as they should.
>>
>
>
> What is the exact problem you do have/see ?

Can you describe in some single repeatable steps a reproducer ?
>

Install udev rules file from mutlipath tools and see that partitions are
not created.


> Which distro are you talking about  (Fedora ? Debian ??)
>

That's Gentoo. But I think any distro that uses whatever upstream prepares
have this problem.


> Have you opened any bugzilla anywhere ?
>

Not yet. I'd like to understand what needs to be done first.

--
Peter.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20160425/d7af8544/attachment.htm>