Project

General

Profile

Actions
Copy link

Bug #5673

closed

mon: upgrade from 0.61.4 to 0.61.5 breaks quorum

Added by Sage Weil almost 11 years ago. Updated almost 11 years ago.

Status:
Won't Fix
Priority:
Immediate
Assignee:
-
Category:
Monitor
Target version:
-
% Done:

0%

Source:
Community (user)
Tags:
Backport:
Regression:
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

maybe the feature bits head an inadvertant change? or the scrub feature bit?

(08:38:56 PM) mikedawson: dmick, sage: after upgrade from 0.61.4 to 0.61.5, restarted my lowest rank mon and get "cephx: verify_reply coudln't decrypt with error: error decoding block for decryption"

lurbs_ saw it as well.

need to reproduce with mon and (pbly) msgr logs

Actions
Copy link
 #1

Updated by Sage Weil almost 11 years ago

  • Status changed from Need More Info to Won't Fix

the problem is that the old mons (0.61.4) are misinterpreting the feature bits of the new mons. the old ones do some aritchmetic which makes them think they require ~0ull, and the new mons now fail to report that that's what they support.

2013-07-18 22:39:49.397378 7f0e3e7ca700 20 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).accept got peer connect_seq 0 global_seq 32
2013-07-18 22:39:49.397409 7f0e3e7ca700 10 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).accept of host_type 1, policy.lossy=0
2013-07-18 22:39:49.397419 7f0e3e7ca700 10 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).accept my proto 10, their proto 10
2013-07-18 22:39:49.397427 7f0e3e7ca700  1 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).peer missing required features fffffffc00000000
2013-07-18 22:39:49.397835 7f0e3e7ca700 10 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).accept couldn't read connect
2013-07-18 22:39:49.397873 7f0e3e7ca700 10 -- 10.214.133.126:6789/0 >> 10.214.133.116:6789/0 pipe(0x28cbc80 sd=24 :6789 s=0 pgs=0 cs=0 l=0).  queued = 0

marking won't fix. updated release notes.

Actions
Copy link

Also available in: Atom PDF