]> git.apps.os.sepia.ceph.com Git - ceph.git/log
ceph.git
16 years agomsgr: clean up refs to static 'rank'
Sage Weil [Fri, 1 May 2009 13:55:54 +0000 (06:55 -0700)]
msgr: clean up refs to static 'rank'

16 years agouclient: clean up mount wrt osdmap, mdsmap
Sage Weil [Thu, 30 Apr 2009 22:24:43 +0000 (15:24 -0700)]
uclient: clean up mount wrt osdmap, mdsmap

16 years agojournaler: store layout in journal header
Sage Weil [Thu, 30 Apr 2009 18:14:31 +0000 (11:14 -0700)]
journaler: store layout in journal header

This means we won't fall over if the 'default' layout changes.  Specify
the layout during journal creation, as you'd expect.

16 years agomds: clean up layouts
Sage Weil [Thu, 30 Apr 2009 17:42:26 +0000 (10:42 -0700)]
mds: clean up layouts

16 years agoosd: more pg_pool, pg changes
Sage Weil [Thu, 30 Apr 2009 16:44:39 +0000 (09:44 -0700)]
osd: more pg_pool, pg changes

Remove 'type' from pg_t.  It's implied based on the pool.

Clean out file_layout.  The pool implies a type and namespace.

16 years agokclient: fix page_mkwrite for 2.6.30
Sage Weil [Thu, 30 Apr 2009 14:12:07 +0000 (07:12 -0700)]
kclient: fix page_mkwrite for 2.6.30

This will change again soon when Nick's other fix gets merged.  Also, we
probably need to stop doing write_begin/end due to some differences in
behavior (e.g., no extending file size? see btrfs as a reference).

16 years agoosd: typo
Sage Weil [Wed, 29 Apr 2009 22:37:39 +0000 (15:37 -0700)]
osd: typo

16 years agoconfig: fix pg pools in default file layouts
Sage Weil [Wed, 29 Apr 2009 19:50:07 +0000 (12:50 -0700)]
config: fix pg pools in default file layouts

16 years agoosd: fix up recovery op accounting
Sage Weil [Wed, 29 Apr 2009 22:33:31 +0000 (15:33 -0700)]
osd: fix up recovery op accounting

Fix varous rop accounting bugs.  Add assertions.  Log recovery
ops.

16 years agotodo
Sage Weil [Wed, 29 Apr 2009 19:06:42 +0000 (12:06 -0700)]
todo

16 years agokclient: retry 2x on ESTALE; fix up mds reply error paths
Sage Weil [Wed, 29 Apr 2009 18:47:13 +0000 (11:47 -0700)]
kclient: retry 2x on ESTALE; fix up mds reply error paths

Also fix error paths, which weren't handling the snap_rwsem
properly.

16 years agokclient: fix symbol names without 'ceph_' prefix
Sage Weil [Wed, 29 Apr 2009 18:44:03 +0000 (11:44 -0700)]
kclient: fix symbol names without 'ceph_' prefix

16 years agokclient: add missing parameter
Yehuda Sadeh [Wed, 29 Apr 2009 19:00:08 +0000 (12:00 -0700)]
kclient: add missing parameter

16 years agoconf: added missing allocations
Yehuda Sadeh [Wed, 29 Apr 2009 18:41:44 +0000 (11:41 -0700)]
conf: added missing allocations

16 years agomon: track per-pg_pool usage
Sage Weil [Wed, 29 Apr 2009 18:38:11 +0000 (11:38 -0700)]
mon: track per-pg_pool usage

16 years agoconfig: make uninit var warning go away
Sage Weil [Wed, 29 Apr 2009 18:13:13 +0000 (11:13 -0700)]
config: make uninit var warning go away

16 years agoosd: push pg_num etc into pg_pool
Sage Weil [Wed, 29 Apr 2009 18:13:01 +0000 (11:13 -0700)]
osd: push pg_num etc into pg_pool

16 years agoosd: clean up assemble_backlog
Sage Weil [Tue, 28 Apr 2009 23:44:12 +0000 (16:44 -0700)]
osd: clean up assemble_backlog

16 years agokclient: don't allow to pass max file size
Yehuda Sadeh [Tue, 28 Apr 2009 23:45:10 +0000 (16:45 -0700)]
kclient: don't allow to pass max file size

16 years agotodo
Sage Weil [Tue, 28 Apr 2009 23:30:34 +0000 (16:30 -0700)]
todo

16 years agoosd: make osd mark itself back up if it is marked down
Sage Weil [Tue, 28 Apr 2009 23:30:12 +0000 (16:30 -0700)]
osd: make osd mark itself back up if it is marked down

We may still want a way to tell an osd to exit.

16 years agomon: optionally dump specific osdmap epoch
Sage Weil [Tue, 28 Apr 2009 23:29:15 +0000 (16:29 -0700)]
mon: optionally dump specific osdmap epoch

16 years agouclient: update dir inode, cap from reply
Sage Weil [Tue, 28 Apr 2009 21:08:57 +0000 (14:08 -0700)]
uclient: update dir inode, cap from reply

16 years agovstart: --nodaemon to run via crun and -f (so that gprof works)
Sage Weil [Tue, 28 Apr 2009 20:42:49 +0000 (13:42 -0700)]
vstart: --nodaemon to run via crun and -f (so that gprof works)

16 years agouclient: take inode ref when dirtying inode
Sage Weil [Tue, 28 Apr 2009 20:22:29 +0000 (13:22 -0700)]
uclient: take inode ref when dirtying inode

16 years agoinitscript: default btrfs mount options notreelog,flushoncommit
Sage Weil [Tue, 28 Apr 2009 19:12:05 +0000 (12:12 -0700)]
initscript: default btrfs mount options notreelog,flushoncommit

16 years agocosd: use \ in dev list in .conf
Sage Weil [Mon, 27 Apr 2009 20:54:23 +0000 (13:54 -0700)]
cosd: use \ in dev list in .conf

16 years agoosd: fix merge_log split point
Sage Weil [Mon, 27 Apr 2009 20:54:10 +0000 (13:54 -0700)]
osd: fix merge_log split point

Look at full eversion, not just .version (part 2).

16 years agomon: order log entries as they're committed
Sage Weil [Tue, 28 Apr 2009 20:10:50 +0000 (13:10 -0700)]
mon: order log entries as they're committed

This doesn't fully order them (across commits), but it helps.

16 years agomon: adjust log levels
Sage Weil [Tue, 28 Apr 2009 05:37:09 +0000 (22:37 -0700)]
mon: adjust log levels

16 years agotodo
Sage Weil [Tue, 28 Apr 2009 05:34:52 +0000 (22:34 -0700)]
todo

16 years agoassert: turn off tls in assert macro
Sage Weil [Tue, 28 Apr 2009 19:54:02 +0000 (12:54 -0700)]
assert: turn off tls in assert macro

About 3% cpu time

16 years agokclient: mark caps sync in check_caps writeback path too
Sage Weil [Tue, 28 Apr 2009 05:16:35 +0000 (22:16 -0700)]
kclient: mark caps sync in check_caps writeback path too

We are already doing so in try_flush_caps().

16 years agokclient: added missing include
Sage Weil [Tue, 28 Apr 2009 05:01:52 +0000 (22:01 -0700)]
kclient: added missing include

16 years agotodo
Sage Weil [Mon, 27 Apr 2009 23:32:43 +0000 (16:32 -0700)]
todo

16 years agoosd: make pg pools resizeable
Sage Weil [Mon, 27 Apr 2009 23:31:50 +0000 (16:31 -0700)]
osd: make pg pools resizeable

Move the pg size from the pgid to the pg_pool descriptor, so that
entire pools of pgs can be resized.

16 years agotodo
Sage Weil [Mon, 27 Apr 2009 20:35:21 +0000 (13:35 -0700)]
todo

16 years agoconfig: -v or --version to print version info
Sage Weil [Mon, 27 Apr 2009 20:32:50 +0000 (13:32 -0700)]
config: -v or --version to print version info

16 years agomon: make down->out osd occur despite mon restarts
Sage Weil [Mon, 27 Apr 2009 19:13:54 +0000 (12:13 -0700)]
mon: make down->out osd occur despite mon restarts

Also add a bit more sanity checking to make sure we don't do the
wrong thing.

16 years agocosd: no valgrind for now
Sage Weil [Mon, 27 Apr 2009 18:53:35 +0000 (11:53 -0700)]
cosd: no valgrind for now

16 years agocommon: add some colors on startup (errors, warnings)
Yehuda Sadeh [Tue, 28 Apr 2009 18:02:55 +0000 (11:02 -0700)]
common: add some colors on startup (errors, warnings)

16 years agoconf: fix allocation
Yehuda Sadeh [Tue, 28 Apr 2009 17:14:05 +0000 (10:14 -0700)]
conf: fix allocation

16 years agoauth: some leaks fixes
Yehuda Sadeh [Mon, 27 Apr 2009 23:26:27 +0000 (16:26 -0700)]
auth: some leaks fixes

16 years agoconf: some more leak fixes
Yehuda Sadeh [Mon, 27 Apr 2009 23:17:53 +0000 (16:17 -0700)]
conf: some more leak fixes

16 years agoconf: some more leak fixes
Yehuda Sadeh [Mon, 27 Apr 2009 21:10:48 +0000 (14:10 -0700)]
conf: some more leak fixes

16 years agocmon: redo remove backslash from end of line
Yehuda Sadeh [Mon, 27 Apr 2009 20:36:17 +0000 (13:36 -0700)]
cmon: redo remove backslash from end of line

16 years agoconf: remove backslash from the end of line
Yehuda Sadeh [Mon, 27 Apr 2009 20:35:26 +0000 (13:35 -0700)]
conf: remove backslash from the end of line

16 years agoconf: add backslash line continuation
Yehuda Sadeh [Mon, 27 Apr 2009 20:31:39 +0000 (13:31 -0700)]
conf: add backslash line continuation

16 years agotodo
Sage Weil [Mon, 27 Apr 2009 18:57:02 +0000 (11:57 -0700)]
todo

16 years agomds: use interval_set for ino allocation
Sage Weil [Mon, 27 Apr 2009 18:52:40 +0000 (11:52 -0700)]
mds: use interval_set for ino allocation

This avoids long ino lists.  Changes ondisk format, unfortunately.

16 years agomon: only log osd boot once
Sage Weil [Mon, 27 Apr 2009 17:50:37 +0000 (10:50 -0700)]
mon: only log osd boot once

16 years agoosd: fix merge_log when split point is divergent
Sage Weil [Mon, 27 Apr 2009 17:22:24 +0000 (10:22 -0700)]
osd: fix merge_log when split point is divergent

Need to compare full eversion (including epoch).

16 years agokclient: avoid spinlock recursion
Yehuda Sadeh [Mon, 27 Apr 2009 18:19:40 +0000 (11:19 -0700)]
kclient: avoid spinlock recursion

16 years agoconf: remove some memory leaks, still have some more
Yehuda Sadeh [Fri, 24 Apr 2009 19:43:30 +0000 (12:43 -0700)]
conf: remove some memory leaks, still have some more

16 years agoupdated sample ceph.conf
Sage Weil [Thu, 23 Apr 2009 23:52:05 +0000 (16:52 -0700)]
updated sample ceph.conf

16 years agoinitscript: cleanup debug
Sage Weil [Thu, 23 Apr 2009 23:48:53 +0000 (16:48 -0700)]
initscript: cleanup debug

16 years agoosd: fix pg missing updates
Sage Weil [Thu, 23 Apr 2009 23:41:34 +0000 (16:41 -0700)]
osd: fix pg missing updates

We may be missing a divergent copy of the same object.

16 years agoosd: do not advance past split point if it is not shared
Sage Weil [Thu, 23 Apr 2009 23:41:09 +0000 (16:41 -0700)]
osd: do not advance past split point if it is not shared

16 years agoinitscript: run valgrind with specified options
Sage Weil [Thu, 23 Apr 2009 22:55:03 +0000 (15:55 -0700)]
initscript: run valgrind with specified options

16 years agoosd: reformat osd map dump
Sage Weil [Thu, 23 Apr 2009 22:53:50 +0000 (15:53 -0700)]
osd: reformat osd map dump

16 years agoosd: set mtime in backlog log entries
Sage Weil [Thu, 23 Apr 2009 22:52:09 +0000 (15:52 -0700)]
osd: set mtime in backlog log entries

16 years agoinitscript: modprobe btrfs and btrfsctl -a before mount
Sage Weil [Thu, 23 Apr 2009 22:12:26 +0000 (15:12 -0700)]
initscript: modprobe btrfs and btrfsctl -a before mount

16 years agomon: fix osd mark out code
Sage Weil [Thu, 23 Apr 2009 22:12:06 +0000 (15:12 -0700)]
mon: fix osd mark out code

16 years agokclient: fix dcache_lock ordering
Sage Weil [Thu, 23 Apr 2009 16:49:13 +0000 (09:49 -0700)]
kclient: fix dcache_lock ordering

16 years agomon: allow clusters of size 2
Sage Weil [Thu, 23 Apr 2009 19:07:56 +0000 (12:07 -0700)]
mon: allow clusters of size 2

16 years agomon: clean up log msg formats a bit
Sage Weil [Thu, 23 Apr 2009 17:41:05 +0000 (10:41 -0700)]
mon: clean up log msg formats a bit

16 years agomon: don't log 'osd alive' updates
Sage Weil [Thu, 23 Apr 2009 17:33:51 +0000 (10:33 -0700)]
mon: don't log 'osd alive' updates

These probably don't need to go in the main log.

16 years agoosd: only trim pg log if pg is complete
Sage Weil [Thu, 23 Apr 2009 17:22:49 +0000 (10:22 -0700)]
osd: only trim pg log if pg is complete

Good policy.  And log.trim() so asserts.

16 years agoosd: set osd down -> out interval to 5 minutes
Sage Weil [Thu, 23 Apr 2009 17:20:11 +0000 (10:20 -0700)]
osd: set osd down -> out interval to 5 minutes

This is the delay from when an osd is marked down before we decide it
won't recover soon as we migrate/replicate data elsewhere.

16 years agoosd: make sure pg_stats_t.reported increases
Sage Weil [Thu, 23 Apr 2009 17:10:51 +0000 (10:10 -0700)]
osd: make sure pg_stats_t.reported increases

Our pg stat updates to the monitor will get (temporarily) ignored if this
jumps backwards.  Make sure we keep the largest value when merging info
from a peer.

16 years agomakefile: add missing Spinlock.h
Sage Weil [Thu, 23 Apr 2009 16:39:29 +0000 (09:39 -0700)]
makefile: add missing Spinlock.h

16 years agomds: make eval_caps safe on non-auth
Sage Weil [Wed, 22 Apr 2009 18:36:11 +0000 (11:36 -0700)]
mds: make eval_caps safe on non-auth

16 years agomds: eval_caps on cap removal
Sage Weil [Wed, 22 Apr 2009 18:35:57 +0000 (11:35 -0700)]
mds: eval_caps on cap removal

16 years agojournaler: fix flush bug
Sage Weil [Wed, 22 Apr 2009 22:55:49 +0000 (15:55 -0700)]
journaler: fix flush bug

Fixes problem with aec1203d8502ad9b8583b53e55491f3a43309b97

Also uses decode for entry sizes (endian-safe!).

16 years agomon: avoid duplicate log entries
Sage Weil [Wed, 22 Apr 2009 22:35:36 +0000 (15:35 -0700)]
mon: avoid duplicate log entries

Also avoid dumping the whole log on 'ceph -w' startup.

16 years agokclient: some sparse fixes
Yehuda Sadeh [Wed, 22 Apr 2009 19:10:24 +0000 (12:10 -0700)]
kclient: some sparse fixes

16 years agokclient: stat -> getattr
Yehuda Sadeh [Wed, 22 Apr 2009 17:47:24 +0000 (10:47 -0700)]
kclient: stat -> getattr

16 years agomds: clean up mds args (rename stat -> getattr)
Sage Weil [Wed, 22 Apr 2009 17:13:21 +0000 (10:13 -0700)]
mds: clean up mds args (rename stat -> getattr)

16 years agotodo
Sage Weil [Tue, 21 Apr 2009 23:48:59 +0000 (16:48 -0700)]
todo

16 years agomds: consolidate loner checks
Sage Weil [Tue, 21 Apr 2009 23:26:15 +0000 (16:26 -0700)]
mds: consolidate loner checks

We want to set the loner before eval'ing locks, and drop loner after
eval'ing (after each has a chance to move out of EXCL).  We also need to
drop loner in each eval so that an individual wrlock drop will do the
right thing.

Still not perfect, but better.

16 years agomds: fix get_caps_issued shift/mask bug
Sage Weil [Tue, 21 Apr 2009 23:24:21 +0000 (16:24 -0700)]
mds: fix get_caps_issued shift/mask bug

16 years agouclient: do local lookups if FILE_RDCACHE on dir and rdcache_gen matches
Sage Weil [Tue, 21 Apr 2009 23:23:42 +0000 (16:23 -0700)]
uclient: do local lookups if FILE_RDCACHE on dir and rdcache_gen matches

16 years agokclient: requeue inode whenever we dirty (use) EXCL|WR caps.
Sage Weil [Tue, 21 Apr 2009 22:12:39 +0000 (15:12 -0700)]
kclient: requeue inode whenever we dirty (use) EXCL|WR caps.

16 years agokclient: queue cap or delayed check if we get unwanted WR|EXCL caps
Sage Weil [Tue, 21 Apr 2009 22:12:06 +0000 (15:12 -0700)]
kclient: queue cap or delayed check if we get unwanted WR|EXCL caps

This ensures we eventually release unwanted WR|EXCL caps.

16 years agokclient: fix delayed counter in check_caps
Sage Weil [Tue, 21 Apr 2009 22:11:25 +0000 (15:11 -0700)]
kclient: fix delayed counter in check_caps

Fix things so we are actually checking whether we delayed a release or
not, so we can act accordingly (by de- or re-scheduling the delayed
cap check).

16 years agojournaler: write out objects as we complete them
Sage Weil [Mon, 20 Apr 2009 23:56:54 +0000 (16:56 -0700)]
journaler: write out objects as we complete them

No reason to delay the write (except maybe to debug log flush()
calls in the MDS).

16 years agotodo
Sage Weil [Mon, 20 Apr 2009 23:23:47 +0000 (16:23 -0700)]
todo

16 years agocosd: conf changes
Sage Weil [Mon, 20 Apr 2009 23:15:46 +0000 (16:15 -0700)]
cosd: conf changes

16 years agomds: finish waiters for root (fixes mds recovery)
Sage Weil [Mon, 20 Apr 2009 23:15:22 +0000 (16:15 -0700)]
mds: finish waiters for root (fixes mds recovery)

16 years agomds: only delete mds on clean shutdown
Sage Weil [Mon, 20 Apr 2009 23:14:52 +0000 (16:14 -0700)]
mds: only delete mds on clean shutdown

This avoids bad asserts on suicide, but still lets us find memory leaks
on clean shutdowns.

16 years agotrivial cleanups
Sage Weil [Mon, 20 Apr 2009 22:54:49 +0000 (15:54 -0700)]
trivial cleanups

16 years agomon: replace mds even if state < 0
Sage Weil [Mon, 20 Apr 2009 22:18:19 +0000 (15:18 -0700)]
mon: replace mds even if state < 0

That conditional test was from before we redid the mds states

16 years agoinit: killall || true
Sage Weil [Mon, 20 Apr 2009 22:17:48 +0000 (15:17 -0700)]
init: killall || true

16 years agoconfig: set g_daemon flag; only create pid file is daemon (even if in foreground)
Sage Weil [Mon, 20 Apr 2009 22:17:13 +0000 (15:17 -0700)]
config: set g_daemon flag; only create pid file is daemon (even if in foreground)

16 years agocosd.conf: more disks
Sage Weil [Mon, 20 Apr 2009 20:34:45 +0000 (13:34 -0700)]
cosd.conf: more disks

16 years agomkcephfs: modprobe btrfs; create mount point earlier
Sage Weil [Mon, 20 Apr 2009 20:34:17 +0000 (13:34 -0700)]
mkcephfs: modprobe btrfs; create mount point earlier

16 years agospinlock: pass 0, not NULL
Sage Weil [Mon, 20 Apr 2009 18:28:08 +0000 (11:28 -0700)]
spinlock: pass 0, not NULL

16 years agokclient: drop i_lock over dput in dcache_readdir
Sage Weil [Mon, 20 Apr 2009 17:43:19 +0000 (10:43 -0700)]
kclient: drop i_lock over dput in dcache_readdir

16 years agokclient: preserve dentry readdir offset during a rename
Sage Weil [Mon, 20 Apr 2009 17:36:46 +0000 (10:36 -0700)]
kclient: preserve dentry readdir offset during a rename

Move the overwritten dentry's readdir offset to the old (renamed,
d_moved) dentry.

16 years agotodo
Sage Weil [Mon, 20 Apr 2009 17:22:04 +0000 (10:22 -0700)]
todo