[linux-lvm] Slow response from lvs and vgs commands

[linux-lvm] Slow response from lvs and vgs commands

[John Griffith]

Hi All,

Wondering if anybody had any info on why lvs and vgs commands may be
taking up to 3 minutes to respond?  In my particular case I'm running
Ubuntu 14.04 as a KVM Instance, with the default version of LVM2
(2.02.98).

There's not anything overly interesting going on with the system, I've
had as few as 1 Volume and as many as 5 when this occurs.  The only
thing that may be unique is that this is during testing so there's a
lot of create/delete activity taking place.  Don't know if that might
have anything to do with it.

I'm wondering if there's something I should be looking at more closely
to debug this, or if there's a way to filter the lvs and vgs commands
to make them more efficient?

Any help or suggestions would be greatly appreciated.

Thanks,
John

Re: [linux-lvm] Slow response from lvs and vgs commands

[Alasdair G Kergon]

On Mon, Jan 19, 2015 at 02:05:39PM -0700, John Griffith wrote:
> I'm wondering if there's something I should be looking at more closely
> to debug this, or if there's a way to filter the lvs and vgs commands
> to make them more efficient?
 
Add -vvvv to the command lines and see if that helps you to see
where the slowness is.

Alasdair

Re: [linux-lvm] Slow response from lvs and vgs commands

[John Griffith]

On Mon, Jan 19, 2015 at 3:21 PM, Alasdair G Kergon <agk@redhat.com> wrote:
> On Mon, Jan 19, 2015 at 02:05:39PM -0700, John Griffith wrote:
>> I'm wondering if there's something I should be looking at more closely
>> to debug this, or if there's a way to filter the lvs and vgs commands
>> to make them more efficient?
>
> Add -vvvv to the command lines and see if that helps you to see
> where the slowness is.
>
> Alasdair
>

Thanks Alasdair,

Turns out I got a great suggestion to wrap this up in strace via the
IRC channel.  Turns out I'm scanning on some devices that are being
plugged in and out of the system and the failure is due to the fact
that the raw device gets ripped out from under the vgscan.

Going to try limiting the devices I interact with via lvm.conf, I
think that will address the issue.

Thanks,
John

Re: [linux-lvm] Slow response from lvs and vgs commands

[Marcos Dutra]

[-- Attachment #1: Type: text/plain, Size: 1481 bytes --]

Hi everybody,

I have the same problem when I use vgs command, so I used vgs -vvv command
and it was slow when the output command showed me this "Archiving volume
group "vgoracle" metadata (seqno 3)", I have many vgs with this status,
Archiving volume group xxx, Is it normal? What can I do to solve this?

Thanks

Marcos Dutra

2015-01-19 20:58 GMT-02:00 John Griffith <john.griffith8@gmail.com>:

> On Mon, Jan 19, 2015 at 3:21 PM, Alasdair G Kergon <agk@redhat.com> wrote:
> > On Mon, Jan 19, 2015 at 02:05:39PM -0700, John Griffith wrote:
> >> I'm wondering if there's something I should be looking at more closely
> >> to debug this, or if there's a way to filter the lvs and vgs commands
> >> to make them more efficient?
> >
> > Add -vvvv to the command lines and see if that helps you to see
> > where the slowness is.
> >
> > Alasdair
> >
>
> Thanks Alasdair,
>
> Turns out I got a great suggestion to wrap this up in strace via the
> IRC channel.  Turns out I'm scanning on some devices that are being
> plugged in and out of the system and the failure is due to the fact
> that the raw device gets ripped out from under the vgscan.
>
> Going to try limiting the devices I interact with via lvm.conf, I
> think that will address the issue.
>
> Thanks,
> John
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@redhat.com
> https://www.redhat.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>

[-- Attachment #2: Type: text/html, Size: 2318 bytes --]

Re: [linux-lvm] Slow response from lvs and vgs commands

[Zdenek Kabelac]

Dne 27.1.2015 v 16:04 Marcos Dutra napsal(a):
> Hi everybody,
>
> I have the same problem when I use vgs command, so I used vgs -vvv command and
> it was slow when the output command showed me this "Archiving volume group
> "vgoracle" metadata (seqno 3)", I have many vgs with this status, Archiving
> volume group xxx, Is it normal? What can I do to solve this?
>

If you are creating lots of randomly named VGs - then you'll need to manual 
prune /etc/lvm/archive directory.

When you use consistently the same VG name - you could configure some limits
via lvm.conf settings:

backup/retain_min
backup/retain_days

(Or you could disable backup completely)

Zdenek