Project

General

Profile

Actions

Bug #123

closed

fix msgr message retry seq numbering

Added by Sage Weil almost 14 years ago. Updated over 13 years ago.

Status:
Resolved
Priority:
High
Assignee:
-
Category:
-
Target version:
% Done:

0%

Source:
Tags:
Backport:
Regression:
Severity:
Reviewed:
Affected Versions:
ceph-qa-suite:
Crash signature (v1):
Crash signature (v2):

Description

we currently assign seq #'s when we send the message over the wire. this numbering breaks when we reconnect because we requeue sent messages but don't adjust out_seq.

the larger issue is that we can revoke messages and pull them from the queue or the sent list, so even sent messages may not be available to resend. what then?

Actions #1

Updated by Sage Weil almost 14 years ago

  • Status changed from New to Resolved

fix:
- we now allow seq #'s to jump forward
- we only assign seq # once after ceph_con_send'ing a message. if it requeues, it goes out again with the original seq.

this is still a stupid protocol. we should just learn how much the peer received during the connection handshake, and pick up from there.. no reason to resend the messages over the wire. but that's a protocol change for later.

Actions

Also available in: Atom PDF