ceph/doc/changelog/v0.80.1.txt

   1 commit a38fe1169b6d2ac98b427334c12d7cf81f809b74 (tag: refs/tags/v0.80.1)
   2 Author: Jenkins <jenkins@inktank.com>
   3 Date:   Mon May 12 15:12:54 2014 +0000
   4
   5     0.80.1
   6
   7 commit ab873622c6d41bc8053a1d205da2e4b55e77351d
   8 Author: Jenkins <jenkins@inktank.com>
   9 Date:   Mon May 12 15:11:33 2014 +0000
  10
  11     0.80.1
  12
  13 commit 03e7d1d550132a55e87972123d5bc3c0bb0cc981
  14 Author: Jenkins <jenkins@inktank.com>
  15 Date:   Mon May 12 15:10:56 2014 +0000
  16
  17     0.80.1
  18
  19 commit 7a4919c7fa10020643df09e012376e4b563ca17d
  20 Author: Jenkins <jenkins@inktank.com>
  21 Date:   Mon May 12 15:09:01 2014 +0000
  22
  23     0.80.1
  24
  25 commit ce852b29dc5cf9857a6bfeef7839293cfadf1877 (refs/remotes/gh/firefly)
  26 Author: Samuel Just <sam.just@inktank.com>
  27 Date:   Fri May 2 16:21:26 2014 -0700
  28
  29     Revert "ReplicatedPG: block scrub on blocked object contexts"
  30
  31     This reverts commit e66f2e36c06ca00c1147f922d3513f56b122a5c0.
  32     Reviewed-by: Sage Weil <sage@inktank.com>
  33
  34     0f3235d46c8fd6c537bd4aa8a3faec6c00f311a8 is the firefly commit
  35     corresponding to e66f2e36c06ca00c1147f922d3513f56b122a5c0.
  36
  37     (cherry picked from commit 84728058dbb91b8ed062240b3373b18078f0c9ca)
  38
  39 commit db8873b69c73b40110bf1512c114e4a0395671ab
  40 Author: Yehuda Sadeh <yehuda@inktank.com>
  41 Date:   Tue May 6 16:55:27 2014 -0700
  42
  43     rgw: fix stripe_size calculation
  44
  45     Fixes: #8299
  46     Backport: firefly
  47     The stripe size calculation was broken, specifically affected cases
  48     where we had manifest that described multiple parts.
  49
  50     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  51     (cherry picked from commit 9968b938b5d47fdf3a67db134bd2ea6bf3a28086)
  52
  53 commit e0fb2e63074d5f71679bb68656336f68e8190f39
  54 Author: Yehuda Sadeh <yehuda@inktank.com>
  55 Date:   Tue May 6 11:06:29 2014 -0700
  56
  57     rgw: cut short object read if a chunk returns error
  58
  59     Fixes: #8289
  60     Backport: firefly, dumpling
  61     When reading an object, if we hit an error when trying to read one of
  62     the rados objects then we should just stop. Otherwise we're just going
  63     to continue reading the rest of the object, and since it can't be sent
  64     back to the client (as we have a hole in the middle), we end up
  65     accumulating everything in memory.
  66
  67     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  68     (cherry picked from commit 03b0d1cfb7bd30a77fedcf75eb06476b21b14e95)
  69
  70 commit 328665db17bf87d36a3dcbf5f0d0d1a9410e2f96
  71 Author: Yehuda Sadeh <yehuda@inktank.com>
  72 Date:   Mon Apr 21 15:34:04 2014 -0700
  73
  74     rgw: send user manifest header field
  75
  76     Fixes: #8170
  77     Backport: firefly
  78     If user manifest header exists (swift) send it as part of the object
  79     header data.
  80
  81     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  82     (cherry picked from commit 5cc5686039a882ad345681133c9c5a4a2c2fd86b)
  83
  84 commit cdb0fac29693fb0250b515891804c4c7fd3045c7
  85 Author: Yan, Zheng <zheng.z.yan@intel.com>
  86 Date:   Fri Apr 11 15:03:37 2014 +0800
  87
  88     client: add asok command to kick sessions that were remote reset
  89
  90     Fixes: #8021
  91     Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
  92     (cherry picked from commit 09a1bc5a4601d356b9cc69be8541e6515d763861)
  93
  94 commit 4a91196bf49936471a731f9316ef904fda982b26
  95 Author: Sage Weil <sage@inktank.com>
  96 Date:   Fri Apr 18 13:50:11 2014 -0700
  97
  98     osd: throttle snap trimmming with simple delay
  99
 100     This is not particularly smart, but it is *a* knob that lets you make
 101     the snap trimmer slow down.  It's a flow and a simple delay, so it is
 102     adjustable at runtime.  Default is 0 (no change in behavior).
 103
 104     Partial solution for #6278.
 105
 106     Signed-off-by: Sage Weil <sage@inktank.com>
 107     (cherry picked from commit 4413670d784efc2392359f0f22bca7c9056188f4)
 108
 109 commit 3ba2e228db1829a02f9cd31225672604fdc8adf7
 110 Author: Sage Weil <sage@inktank.com>
 111 Date:   Fri May 2 14:48:35 2014 -0700
 112
 113     mon/MonClient: remove stray _finish_hunting() calls
 114
 115     Callig _finish_hunting() clears out the bool hunting flag, which means we
 116     don't retry by connection to another mon periodically.  Instead, we send
 117     keepalives every 10s.  But, since we aren't yet in state HAVE_SESSION, we
 118     don't check that the keepalives are getting responses.  This means that an
 119     ill-timed connection reset (say, after we get a MonMap, but before we
 120     finish authenticating) can drop the monc into a black hole that does not
 121     retry.
 122
 123     Instead, we should *only* call _finish_hunting() when we complete the
 124     authentication handshake.
 125
 126     Fixes: #8278
 127     Backport: firefly, dumpling
 128     Signed-off-by: Sage Weil <sage@inktank.com>
 129     Reviewed-by: Joao Eduardo Luis <joao.luis@inktank.com>
 130     (cherry picked from commit 77a6f0aefebebf057f02bfb95c088a30ed93c53f)
 131
 132 commit 383f6440ce34b0d84c0dfa06f561896dad6862f4
 133 Author: Sage Weil <sage@inktank.com>
 134 Date:   Fri May 2 16:41:26 2014 -0700
 135
 136     osd/ReplicatedPG: fix trim of in-flight hit_sets
 137
 138     We normally need to stat the hit_set to know how many bytes to adjust the
 139     stats by.  If the hit_set was just written, we will get ENOENT.
 140
 141     Get the obc instead, which will either get the in-memory copy (because the
 142     repop is still in flight) or load it off of disk.
 143
 144     Fixes: #8283
 145     Backport: firefly
 146     Signed-off-by: Sage Weil <sage@inktank.com>
 147     (cherry picked from commit 72fdd557c35cb721d4b502c5a8f68c878f11a19c)
 148
 149 commit ef35448e63916102261ff6198f2c0ab4e4a57e9b
 150 Author: Sage Weil <sage@inktank.com>
 151 Date:   Tue May 6 11:01:27 2014 -0700
 152
 153     osd/ReplicatedPG: fix whiteouts for other cache mode
 154
 155     We were special casing WRITEBACK mode for handling whiteouts; this needs to
 156     also include the FORWARD and READONLY modes.  To avoid having to list
 157     specific cache modes, though, just check != NONE.
 158
 159     Fixes: #8296
 160     Backport: firefly
 161     Signed-off-by: Sage Weil <sage@inktank.com>
 162     (cherry picked from commit 3e387d62ed95898db8a7d7163c2bacc363b8f617)
 163
 164 commit e7df73dd7aaf5a0b1171f73d6695d26cd25b7b35
 165 Author: Sage Weil <sage@inktank.com>
 166 Date:   Thu May 1 16:53:17 2014 -0700
 167
 168     osd: Prevent divide by zero in agent_choose_mode()
 169
 170     Fixes: #8175
 171     Backport: firefly
 172
 173     Signed-off-by: David Zafman <david.zafman@inktank.com>
 174     Signed-off-by: Sage Weil <sage@inktank.com>
 175     (cherry picked from commit f47f867952e6b2a16a296c82bb9b585b21cde6c8)
 176
 177 commit 022d467b5d6b77c17b6fdaeec8369cae61e9e5a4
 178 Author: David Zafman <david.zafman@inktank.com>
 179 Date:   Mon Apr 21 23:52:04 2014 -0700
 180
 181     osd, common: If agent_work() finds no objs to work on delay 5 (default) secs
 182
 183     Add config osd_agent_delay_time of 5 seconds
 184     Honor delay by ignoring agent_choose_mode() calls
 185     Add tier_delay to logger
 186     Treat restart after delay like we were previously idle
 187
 188     Fixes: #8113
 189     Backport: firefly
 190
 191     Signed-off-by: David Zafman <david.zafman@inktank.com>
 192     (cherry picked from commit b7d31e5f5952c631dd4172bcb825e77a13fc60bc)
 193
 194 commit 6a55c3bc3caf46652e962fa9434900fb494d1e6c
 195 Author: David Zafman <david.zafman@inktank.com>
 196 Date:   Thu May 1 18:54:30 2014 -0700
 197
 198     osd/ReplicatedPG: agent_work() fix next if finished early due to start_max
 199
 200     Backport: firefly
 201
 202     Signed-off-by: David Zafman <david.zafman@inktank.com>
 203     (cherry picked from commit 9cf470cac8dd4d8f769e768f2de6b9eb67a3c3af)
 204
 205 commit 1b899148a729235ab2835d368077f18e62a36a93
 206 Author: Haomai Wang <haomaiwang@gmail.com>
 207 Date:   Sat May 3 12:53:06 2014 +0800
 208
 209     Fix clone problem
 210
 211     When clone happened, the origin header also will be updated in GenericObjectMap,
 212     so the new header wraper(StripObjectHeader) should be updated too.
 213
 214     Fix #8282
 215     Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
 216     (cherry picked from commit 3aee1e0ffe0583f74c02d9c9e86c7fb267f3515c)