luminous: msg/async: do not bump connect_seq for fault during ACCEPTING_SESSION

[chenerqi]

backport tracker issue: https://tracker.ceph.com/issues/42318

---

partial backport (1 commit) of: #24546

parent tracker: https://tracker.ceph.com/issues/42316

[neha-ojha]

Is there a tracker issue for this backport?

[chenerqi]

Is there a tracker issue for this backport?

No, I met the issue in my ceph cluster, it needs fix for ACCEPTING phase fault, this part of code has been rewritten in which Sage fixed the problem in msg/async/ProtocolV2 with commit b2b1234
I didn't find track issue for Sage's commit.

There is a similar track issue for CONNECTING phase fault, Sage fixed it in msg/async/ProtocolV1
#25343
https://tracker.ceph.com/issues/36612
36612 links to luminous backport issue 37521, but I checked luminous branch has no this issue, so this track issue can be closed since nothing to backport to luminous for CONNECTING phase fault.
https://tracker.ceph.com/issues/37521

[smithfarm]

@chenerqi The corresponding master PR for this backport is #24546

It was merged for nautilus. That means: if it is going to be backported, there needs to be a tracker issue. The tracker issue should be created with the following fields:

Status: Pending Backport
Backport: mimic, luminous
Pull request ID: 24546

(In the description just briefly explain the bug that #24546 is fixing)

Once that tracker is created, we can proceed. Adding DNM until then.

[chenerqi]

@chenerqi The corresponding master PR for this backport is #24546

It was merged for nautilus. That means: if it is going to be backported, there needs to be a tracker issue. The tracker issue should be created with the following fields:

Status: Pending Backport
Backport: mimic, luminous
Pull request ID: 24546

(In the description just briefly explain the bug that #24546 is fixing)

Once that tracker is created, we can proceed. Adding DNM until then.

Tracker issue 42316 created.
https://tracker.ceph.com/issues/42316

[smithfarm]

@chenerqi Thanks. There's one more thing left to do: fix the commit message. The correct commit message should be:

msg/async/ProtocolV2: do not bump connect_seq for fault during ACCEPTING_SESSION

If we have a connection race, and we lose, we may end up with outgoing
messages *and* be in ACCEPTING_SESSION.  If we then fault, we want to
leave connect_seq at 0 to avoid triggering a reset.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b2b123435f32338826afc4a662e64c29f1262f68)

Conflicts:
     src/msg/async/ProtocolV2.cc
- file does not exist in luminous. Made change manually in the file src/msg/async/AsyncConnection.cc

(in other words, please run git commit --amend and replace the existing commit message with the text shown above)

[smithfarm]

commit message is wrong

[smithfarm]

@rjfd Does this look right to you for luminous?

[chenerqi]

@chenerqi Thanks. There's one more thing left to do: fix the commit message. The correct commit message should be:

msg/async/ProtocolV2: do not bump connect_seq for fault during ACCEPTING_SESSION

If we have a connection race, and we lose, we may end up with outgoing
messages *and* be in ACCEPTING_SESSION.  If we then fault, we want to
leave connect_seq at 0 to avoid triggering a reset.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b2b123435f32338826afc4a662e64c29f1262f68)

Conflicts:
     src/msg/async/ProtocolV2.cc
- file does not exist in luminous. Made change manually in the file src/msg/async/AsyncConnection.cc

(in other words, please run git commit --amend and replace the existing commit message with the text shown above)

Updated.

commit message looks right, now