[DRBD-user] Resync network error

[Josep Maria Viñolas]

Hi,

We are experiencing a resync problem with drbd dkms kernel module 9.0.25-1:

# dkms status
> drbd, 9.0.25-1, 4.19.0-12-amd64, x86_64: installed
>

If we do not set any parameters in global_common.conf seems to be ok but
extremely low resync performance noticed (few kbits/s). When we set any
resync parameter (and drbdadm adjust all on both nodes) we can't reach more
than 5/6Mbps in syncing nodes (both with same nvme high speed disks) and
this error message is shown in dmesg:

[  851.911236] drbd storage/0 drbd1000 cluster1: ASSERTION
> __dec_unacked(peer_device) >= 0 FAILED in e_end_resync_block
> [  853.771264] drbd storage/0 drbd1000 cluster1: ASSERTION
> __dec_unacked(peer_device) >= 0 FAILED in e_end_resync_block
> [  855.683292] drbd storage/0 drbd1000 cluster1: ASSERTION
> __dec_unacked(peer_device) >= 0 FAILED in e_end_resync_block
> [  855.807294] drbd storage/0 drbd1000 cluster1: ASSERTION
> __dec_unacked(peer_device) >= 0 FAILED in e_end_resync_block
>

The net parameters  we set in global_common.conf are:

> net {
> protocol C;
>
> allow-two-primaries;
> after-sb-0pri discard-zero-changes;
> after-sb-1pri discard-secondary;
> after-sb-2pri disconnect;
> }
>

And we have MTU 9000 on both nodes.

Any ideas why this is happening?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20201123/83ea3b2e/attachment.htm>