1 commit 59ddece17e36fef69ecf40e239aeffad33c9db35
2 Author: Gary Lowell <gary.lowell@inktank.com>
3 Date: Tue Jul 23 13:52:19 2013 -0700
7 commit 38c3271d3fc415919f0856398bd94eb87a0776b5
8 Author: Sage Weil <sage@inktank.com>
9 Date: Tue Jul 23 13:32:12 2013 -0700
11 mon/OSDMonitor: fix base case for 7fb3804fb workaround
13 After cluster creation, we have no full map stored and first_committed ==
14 1. In that case, there is no need for a full map, since we can get there
15 from OSDMap() and the incrementals.
18 Signed-off-by: Sage Weil <sage@inktank.com>
19 Reviewed-by: Joao Eduardo Luis <joao@inktank.com>
20 (cherry picked from commit e807770784175b05130bba938113fdbf874f152e)
22 commit f94956cb1a56ff62e01b7ae218a93c4004470ae5
23 Author: Joao Eduardo Luis <joao.luis@inktank.com>
24 Date: Tue Jul 23 17:25:13 2013 +0100
26 mon: OSDMonitor: work around a full version bug introduced in 7fb3804fb
28 In 7fb3804fb860dcd0340dd3f7c39eec4315f8e4b6 we moved the full version
29 stashing logic to the encode_trim_extra() function. However, we forgot
30 to update the osdmap's 'latest_full' key that should always point to
31 the latest osdmap full version. This eventually degenerated in a missing
32 full version after a trim. This patch works around this bug by looking
33 for the latest available full osdmap version in the store and updating
34 'latest_full' to its proper value.
39 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
40 Reviewed-by: Sage Weil <sage@inktank.com>
41 (cherry picked from commit 97462a3213e5e15812c79afc0f54d697b6c498b1)
43 commit 10e1de45dc8ace793ecf921f884f90c9daa99c48
44 Author: Joao Eduardo Luis <joao.luis@inktank.com>
45 Date: Tue Jul 23 16:36:52 2013 +0100
47 mon: OSDMonitor: update the osdmap's latest_full with the new full version
49 We used to do this on encode_full(), but since [1] we no longer rely on
50 PaxosService to manage the full maps for us. And we forgot to write down
51 the latest_full version to the store, leaving it in a truly outdated state.
53 [1] - 7fb3804fb860dcd0340dd3f7c39eec4315f8e4b6
57 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
58 Reviewed-by: Sage Weil <sage@inktank.com>
59 (cherry picked from commit a815547ed3e5ffdbbb96c8c0c1b8d6dd8c62bfba)
61 commit a0cb40b45c4f2f921a63c2d7bb5a28572381d793
62 Author: Sage Weil <sage@inktank.com>
63 Date: Thu Jul 18 14:35:19 2013 -0700
65 mon: decline to scrub when paxos is not active
67 In f1ce8d7c955a2443111bf7d9e16b4c563d445712 we close a race between scrub
68 and paxos commit completion on the leader. The fix is nontrivial to
69 backport and probably not worthwhile; just avoid scrubbing at that time
72 Note that the actual fix for this is in commit
73 f1ce8d7c955a2443111bf7d9e16b4c563d445712.
75 Signed-off-by: Sage Weil <sage@inktank.com>