Commit Graph

13851 Commits

Author SHA1 Message Date
Sage Weil
9c731eddfb mdsmap: initialize standby_for_rank
This is initialized in MDSMonitor anyway; do so where for completeness.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 14:14:14 -07:00
Sage Weil
5e27a079e8 mon: simplify mds follow checks
Instead of assigning followers in the last_beacon laggy check loop, do it
at the end, the same way we let standby nodes take over.

This also fixes a bug where a non-standby node (say, up:replay) that used
to be up:standby-replay and has standby_for_rank set gets reset back to
up:standby-replay.

Fixes: #1001
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 14:13:56 -07:00
Sage Weil
2b59bc6192 mon: simplify mds laggy check
We should never have a laggy standby, so technically this doesn't change
any behavior, but it makes the flow less confusing.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 14:08:43 -07:00
Sage Weil
e9afe17e85 mon: don't take over for a standby-replay
If a standby-replay is laggy we shouldn't "take over" for them (they're
not part of the cluster yet).  They should be removed like a regular
standby.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 14:08:40 -07:00
Colin Patrick McCabe
2b2766dd46 rados-tool: use init_with_config interface
Programs that use both librados and common_init should use
init_with_config.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-12 14:04:06 -07:00
Sage Weil
bc98fe7e42 mds: improve scatterlog debug msg 2011-04-12 11:12:56 -07:00
Sage Weil
c7f9d2f9a8 mds: clear flush state on rejoin ack
If we sent scatterlock state during rejoin, the auth will send us an inode
base.  Clear scatterlock flush state if that happens.

Fixes: #637
Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-12 11:12:56 -07:00
Sage Weil
c9872a44a3 mds: send any dirty scatterlock state on rejoin
Not just inodes for auth dirfrags, but for any inode with dirty scatterlock
state.  Include the root inode.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-12 11:12:56 -07:00
Sage Weil
43d74721c6 mds: reset dirty->flushing on rejoin scatterflush
Reset dirty/flushing state during rejoin.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-12 11:12:56 -07:00
Sage Weil
2a1af8b86e mds: drop debugging cruft
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 11:12:19 -07:00
Sage Weil
0f10e95498 mds: make _create_system_file dirty dentries properly
Properly dirty the new dentries so they get written to the directory
objects later on.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 11:12:16 -07:00
Sage Weil
31fc9edb81 mds: fix create_mydir_hierarchy to save dir
Mark the dentries dirty so they get saved to disk (they're not journaled!).
This fixes rstat problems on startup, where populate_mydir was recreating
the entries and munging rstats accordingly.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-12 11:12:12 -07:00
Colin Patrick McCabe
f4d86777db radosacl: read Ceph configuration file
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-12 10:55:48 -07:00
Colin Patrick McCabe
ceba04f4dd cephtool: don't print anything on SIGTERM, SIGINT
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-12 09:40:48 -07:00
Sage Weil
21df04276c mds: don't crash on bad export_dir command
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-11 21:09:28 -07:00
Sage Weil
fb8c09b0d7 mds: fix file_eval on subtree roots
We were going from mix->sync when we shouldn't.  Fix mis-coded filelock
and look at _lock_ state, not inode's!

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-11 21:09:24 -07:00
Sage Weil
1f23a22992 Merge remote branch 'origin/dead-code' 2011-04-11 16:30:10 -07:00
Sage Weil
7ca69ebe99 osd: fix ref leak on pg when queueing
op_wq.queue() takes a ref; don't take one here too.  We were leaking like
crazy.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-11 16:24:57 -07:00
Sage Weil
6768586cf5 osd: take reference for RepModify
We put in the completions on done; take the matching reference.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-11 16:24:57 -07:00
Sage Weil
60b645e5a5 osd: fix race in op enqueuing
This is never really exposed except when op_threads == 0.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-11 16:24:57 -07:00
Sage Weil
cc5dd6d6e1 mon: simplify osd health output
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-11 09:43:25 -07:00
Sage Weil
9c0d42fe99 osd: only log scrub errors (not ok's)
Otherwise we fill up the error log with "foo is ok!" messages.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-11 09:36:25 -07:00
Samuel Just
eacfb8d381 TestSnaps: Fix compile error
Introduced in 0e371de79e.

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
2011-04-11 09:45:57 -07:00
Colin Patrick McCabe
c02b56e622 cfuse: pass through SIGTERM, SIGINT to FUSE
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-08 18:16:58 -07:00
Colin Patrick McCabe
1cb4aae45a ceph_fuse_ll_main: restructure error handling
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-08 18:16:58 -07:00
Josh Durgin
b3f5102d68 librbdtest: check for error on out of bounds I/O
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
2011-04-08 17:39:11 -07:00
Josh Durgin
46c761fa93 librbd: check for out of bounds I/O in all forms of read/write
This was found by qemu-io tests, which tried to read and write past
the end of an rbd image. The test hung waiting for a completion that
was never scheduled, since it did not check the return value of
rbd_aio_write.

Reported-by: Stefan Hajnoczi <stefanha@gmail.com>
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
2011-04-08 17:39:11 -07:00
Brian Chrisman
e34078b2ed cfuse recognizing same xattr subset as kernel client
Hopefully this patch is formatted/signed off correctly.

Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
2011-04-08 16:12:56 -07:00
Samuel Just
0e371de79e TestSnaps.cc: allow client id to be overridden
CEPH_CLIENT_ID environment variable if set now overrides the rados
client id for testsnaps

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
2011-04-08 16:15:19 -07:00
Sage Weil
2b0e474a0a rgw: utility, not daemon
- we don't want to open a log file
- we don't want to print a stability warning
- we don't want to fork (not that we do, but anyway!)

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-04-08 13:11:45 -07:00
Greg Farnum
7f3a054029 CDir: check_rstats will now print out dir stats whenever there's a bug.
Previously it only printed out dir stats at high debug levels, which
meant you could get output of the dentries without seeing what the
totals were supposed to be!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
2011-04-08 11:17:48 -07:00
Colin Patrick McCabe
f3f7383ce9 obsync: preserve ACLs
Preserve ACLs by default. Also add --no-preserve-acls option to disable
this behavior.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 20:07:44 -07:00
Colin Patrick McCabe
556f06e8d1 obsync: escape obj names between S3 and local FS
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 16:09:59 -07:00
Greg Farnum
4995117325 osd: futz with stat again.
We really need to fill in the result, even if it's that the object
doesn't exist!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
2011-04-07 16:30:26 -07:00
Samuel Just
0de233c386 test/osd: Add latency stat collection to TestSnaps.cc/RadosModel.h
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
2011-04-07 16:08:05 -07:00
Samuel Just
e26626a8b4 ReplicatedPG: fix error in snap_trimmer
newsnaps continued to be used after it was swapped with coi.snaps.

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
2011-04-07 16:07:30 -07:00
Tommi Virtanen
84aaa6e08b crush, clitest: Port old test_crushtool.sh to clitest.
Also test --help, while we're at it.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 15:05:02 -07:00
Sage Weil
61ce52cc33 osd: more futzing with stat
We can get here when the object doesn't exist if the client specifies
may read and may write (in this case, Filer::probe).  Look at the exists
bit in the object context.  If the object is supposed to exist, we should never get ENOENT from stat.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-07 15:01:58 -07:00
Tommi Virtanen
eb08daaba2 libceph: Make header usable from plain old C.
Fixes: #986
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 14:53:45 -07:00
Tommi Virtanen
eee1f920e4 crush: Remove dead code.
The Makefile never integrated with the surrounding automake
setup, and is unused.

The test helper has failed to compile for quite a long time.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 14:42:05 -07:00
Tommi Virtanen
aefa10464f crush: Remove dead code.
Looks like these never went anywhere.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 14:34:37 -07:00
Wido den Hollander
38c6831c38 radosgw: Remove PHP setting from VirtualHost documentation
Signed-off-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Sage Weil <sage@newdream.net>
2011-04-07 14:32:31 -07:00
Tommi Virtanen
05596d7e62 osd: Remove dead code.
RAID4PG has been "temporarily ripped out" for 3.5 years.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 14:25:12 -07:00
Tommi Virtanen
f81b505c37 msg, os: Remove dead code.
Looks like these never went anywhere.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
2011-04-07 14:22:39 -07:00
Colin Patrick McCabe
033e29a49d test-obsync.py: add test for --follow-symlinks
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 14:22:30 -07:00
Colin Patrick McCabe
2506819d90 test-obsync: some fixes, be more verbose
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 14:22:30 -07:00
Colin Patrick McCabe
2653361007 boto_tool.py: fix rmbucket
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 14:22:30 -07:00
Colin Patrick McCabe
42975b12de obsync: add --follow-symlinks
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-04-07 14:22:30 -07:00
Greg Farnum
9599b66fba ReplicatedPG: tolerate ENOENT when statting.
This got broken in commit f842bcf893
so that on any MDS restart the OSD would die. Bad OSD! Bad!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
2011-04-07 14:37:23 -07:00
Tommi Virtanen
a66da4af25 qa: Remove fsx workunit.
Covered better by autotest, these days.
2011-04-07 14:07:24 -07:00