Bogus "sha1" resource

[elonen]

On one of my LINSTOR cluster nodes, drbdtop shows a non-existing resource "sha1" with no nodes. It only shows up in drbdtop listing, not in /dev/drbd/by-res, drbd-overview or /var/lib/linstor.d/.

DRBDTOP v0.2.1-3-g70075d0 (kernel: v9.0.16/utils: v9.5.0)
...
│  sha1                 | Secondary | ✓     | ✓          | ✓           | ✓       | ✓

DRBDTOP v0.2.1-3-g70075d0 (kernel: v9.0.16/utils: v9.5.0) - Details for sha1
┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│Resource: sha1: (Overall danger score: 0)                                                                                                                           │
│ Local Disc(Secondary):                                                                                                                                             │
│                                                                                                                                                                    │
│

[deleteriousEffect]

That's really interesting!

Could you provide me with the output of drbdsetup events2 --now --verbose --statistics on that machine?

There's also a /etc/drbd.d directory that might be of interest to us here.

---

From: elonen notifications@github.com Sent: Sunday, October 28, 2018 1:21 PM To: LINBIT/drbdtop Cc: Subscribed Subject: [LINBIT/drbdtop] Bogus "sha1" resource (#11)

On one of my LINSTOR cluster nodes, drbdtop shows a non-existing resource "sha1" with no nodes. It only shows up in drbdtop listing, not in /dev/drbd/by-res, drbd-overview or /var/lib/linstor.d/.

DRBDTOP v0.2.1-3-g70075d0 (kernel: v9.0.16/utils: v9.5.0)

...

│ sha1 | Secondary | ✓ | ✓ | ✓ | ✓ | ✓

DRBDTOP v0.2.1-3-g70075d0 (kernel: v9.0.16/utils: v9.5.0) - Details for sha1

┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐

│Resource: sha1: (Overall danger score: 0) │

│ Local Disc(Secondary): │

│ │

│

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[elonen]

Ok, there it is in drbdsetup output:

exists peer-device name:vm-103-disk-1 peer-node-id:1 conn-name:mox-a volume:0 replication:Established peer-disk:UpToDate peer-client:no resync-suspended:no received:1
087504 sent:0 out-of-sync:0 pending:0 unacked:0
exists peer-device name:vm-103-disk-1 peer-node-id:0 conn-name:mox-b volume:0 replication:Established peer-disk:UpToDate peer-client:no resync-suspended:no received:8
48 sent:0 out-of-sync:0 pending:0 unacked:0
exists resource name:sha1 role:Secondary suspended:no write-ordering:none
exists resource name:vm-104-disk-1 role:Secondary suspended:no write-ordering:flush
exists connection name:vm-104-disk-1 peer-node-id:0 conn-name:mox-a connection:Connected role:Secondary congested:no ap-in-flight:0 rs-in-flight:0
exists connection name:vm-104-disk-1 peer-node-id:1 conn-name:mox-b connection:Connected role:Secondary congested:no ap-in-flight:0 rs-in-flight:0

Not present in /etc/drbd.d.

[rck]

Did that resource ever exist? Like did you create it with linstor at one point? drbdsetup status sha1 would also be interesting.

[elonen]

No, never explicitly created a resource names sha1. My best guess is a misplaced command argument somewhere in the chain (I'm using LINSTOR Proxmox plugin and doing manual maintenance on drbdadmn level, so there are a lot of layers where this might have happened).

The resource only shows up on one node, and status report is rather short:

# drbdsetup status sha1
sha1 role:Secondary

[rck]

Hm, all in all most likely not a bug in drbdtop. It is strange that it matches an actual value that is set in the res files LINSTOR creates. AFAIR it sets cram-hmac-alg sha1;. To produce such a state as yours, a drbdsetup new-resource sha1 0 is enough, but who/and why would that have been executed?

Any other hints for that? Syslogs? Proxmox logs? Shell history? If you just want to get rid of it, you can issue a drbdsetup down sha1. Then it would be interesting if drbdadm up sha1 regenerates it. I assume it does not, because then there needs to be a res file and we have to find it :)

[elonen]

Reboot and/or package upgrade seems to have exorcised it, and drbdadm up sha1 doesn't bring it back on any node. Guess we'll never know. :)