Mailing List Archive: thin volume looses sparse during replication

thin volume looses sparse during replication

Oct 24, 2021, 3:43 AM

Post #1 of 3 (402 views)

Hello

when performing the initial synchronization of an thin-provisioned LVM2
volume from primary to secondary, I can see data gets written on
secondary and the resulting LVM2 volume, although supposedly "thin",
ends up with 100% used data.

To fix that, I need to `blkdiscard` it (even from the primary, that
works too) after synchronization. And this is an extraordinary painful
step to go through. Because of that limitation, I cannot create many
drbd resources that would over-commit physically available space at
once: I need to sync/discard the resources step by step...

Is there a way to make the initial replication keep being thin (or
sparse, as in file-system terminology, when dealing with zeroes)?

Thank you!

Best regards,
--
Pierre-Philipp Braun
SMTP Health Campaign: enforce STARTTLS and verify MX certificates
<https://nethence.com/smtp/>
_______________________________________________
Star us on GITHUB: https://github.com/LINBIT
drbd-user mailing list
drbd-user@lists.linbit.com
https://lists.linbit.com/mailman/listinfo/drbd-user

Re: thin volume looses sparse during replication [ In reply to ]

gianni.milo22 at gmail

Oct 24, 2021, 7:25 AM

Post #2 of 3 (402 views)

Permalink

Check the section in DRBD user guide, "4.1.8. Skipping initial
resynchronization". I think that's what you are looking for, assuming that
there's no need to preserve the data on existing volumes (i.e you start
with blank LVM volumes).

On Sun, 24 Oct 2021 at 11:55, Pierre-Philipp Braun <pbraun@nethence.com>
wrote:

> Hello
>
> when performing the initial synchronization of an thin-provisioned LVM2
> volume from primary to secondary, I can see data gets written on
> secondary and the resulting LVM2 volume, although supposedly "thin",
> ends up with 100% used data.
>
> To fix that, I need to `blkdiscard` it (even from the primary, that
> works too) after synchronization. And this is an extraordinary painful
> step to go through. Because of that limitation, I cannot create many
> drbd resources that would over-commit physically available space at
> once: I need to sync/discard the resources step by step...
>
> Is there a way to make the initial replication keep being thin (or
> sparse, as in file-system terminology, when dealing with zeroes)?
>
> Thank you!
>
> Best regards,
> --
> Pierre-Philipp Braun
> SMTP Health Campaign: enforce STARTTLS and verify MX certificates
> <https://nethence.com/smtp/>
> _______________________________________________
> Star us on GITHUB: https://github.com/LINBIT
> drbd-user mailing list
> drbd-user@lists.linbit.com
> https://lists.linbit.com/mailman/listinfo/drbd-user
>

Re: thin volume looses sparse during replication [ In reply to ]

pbraun at nethence

Nov 1, 2021, 1:06 AM

Post #3 of 3 (396 views)

Permalink

> Check the section in DRBD user guide, "4.1.8. Skipping initial
> resynchronization". I think that's what you are looking for, assuming
> that there's no need to preserve the data on existing volumes (i.e you
> start with blank LVM volumes).

Thanks, that helped a lot.

--
Pierre-Philipp Braun
SMTP Health Campaign: enforce STARTTLS and verify MX certificates
<https://nethence.com/smtp/>
_______________________________________________
Star us on GITHUB: https://github.com/LINBIT
drbd-user mailing list
drbd-user@lists.linbit.com
https://lists.linbit.com/mailman/listinfo/drbd-user