]> git.apps.os.sepia.ceph.com Git - ceph.git/log
ceph.git
16 years agokclient: script to build patch set for upstream
Sage Weil [Mon, 20 Oct 2008 16:24:52 +0000 (09:24 -0700)]
kclient: script to build patch set for upstream

16 years agokclient: some mon client cleanup
Sage Weil [Mon, 20 Oct 2008 16:24:25 +0000 (09:24 -0700)]
kclient: some mon client cleanup

16 years agokclient: kbuild patch update
Sage Weil [Sun, 19 Oct 2008 20:43:56 +0000 (13:43 -0700)]
kclient: kbuild patch update

16 years agokclient: more messenger cleanups, comments
Sage Weil [Sat, 18 Oct 2008 04:12:52 +0000 (21:12 -0700)]
kclient: more messenger cleanups, comments

16 years agokclient: messenger.c cleanup, reorganization, comments
Sage Weil [Sat, 18 Oct 2008 03:47:07 +0000 (20:47 -0700)]
kclient: messenger.c cleanup, reorganization, comments

16 years agokclient: fix cap_snap flush to reference correct snap
Sage Weil [Fri, 17 Oct 2008 21:12:35 +0000 (14:12 -0700)]
kclient: fix cap_snap flush to reference correct snap

The client was passing the snapid the data was to fall into
instead of the snap it _follows_ (which is now the mds tends
to look at things).  Use snapid-1 instead.  This fixes the
problem of files being written while the snapshot was taken
appearing as 0 bytes in the snapshot.

16 years agoosd todos
Sage Weil [Fri, 17 Oct 2008 20:36:15 +0000 (13:36 -0700)]
osd todos

16 years agomds: only check osdmap against snap table while active
Sage Weil [Fri, 17 Oct 2008 20:30:33 +0000 (13:30 -0700)]
mds: only check osdmap against snap table while active

16 years agoosd: track snap collections per pg, and clean them up
Sage Weil [Fri, 17 Oct 2008 20:27:11 +0000 (13:27 -0700)]
osd: track snap collections per pg, and clean them up

Explicitly track which snap collections exist for each pg.  Clean them
up when the PG is destroyed.

16 years agoosd: fast path if snap collection doesn't exist in snap_trimmer
Sage Weil [Fri, 17 Oct 2008 19:55:23 +0000 (12:55 -0700)]
osd: fast path if snap collection doesn't exist in snap_trimmer

16 years agofilestore: return error code (instead of asserting) if collection doesn't exist on...
Sage Weil [Fri, 17 Oct 2008 19:53:46 +0000 (12:53 -0700)]
filestore: return error code (instead of asserting) if collection doesn't exist on collection_list

The OSD doesn't pre-create the snap collections.

16 years agovstartnew.sh: turn up debugging on osd
Sage Weil [Fri, 17 Oct 2008 19:44:17 +0000 (12:44 -0700)]
vstartnew.sh: turn up debugging on osd

16 years agomds: cleanly update snapserver table after snap purge
Sage Weil [Fri, 17 Oct 2008 19:43:15 +0000 (12:43 -0700)]
mds: cleanly update snapserver table after snap purge

Any modifications to the table need to be journaled.  Create MDSTableServer infrastructure
for a simple update done by the server (no prepare/commit), and use it to
remove the snapid(s).  Replay no works.

16 years agokclient: more snap.c comments, cleanup
Sage Weil [Fri, 17 Oct 2008 19:17:28 +0000 (12:17 -0700)]
kclient: more snap.c comments, cleanup

16 years agokclient: a few more snap.c comments
Sage Weil [Fri, 17 Oct 2008 17:17:13 +0000 (10:17 -0700)]
kclient: a few more snap.c comments

16 years agokclient: adjust ceph_finish_cap behavior
Sage Weil [Fri, 17 Oct 2008 14:26:03 +0000 (07:26 -0700)]
kclient: adjust ceph_finish_cap behavior

Look at capsnap->dirty instead of used cap refs, as this directly indicates
whether _this_ capsnap can be flushed (used bit indicates whether _all_
capsnaps can be flushed.. no need to wait that long!).  Avoid an extra
call to ceph_flush_snaps by waiting until after to call it.

16 years agokclient: some snap.c comments, cleanup
Sage Weil [Fri, 17 Oct 2008 14:12:04 +0000 (07:12 -0700)]
kclient: some snap.c comments, cleanup

16 years agokclient: osd messages timeout retriggering
Yehuda Sadeh [Fri, 17 Oct 2008 19:31:00 +0000 (12:31 -0700)]
kclient: osd messages timeout retriggering

16 years agokclient: refine osd messages timeout handling
Yehuda Sadeh [Fri, 17 Oct 2008 16:58:25 +0000 (09:58 -0700)]
kclient: refine osd messages timeout handling

16 years agokclient: debug message
Yehuda Sadeh [Thu, 16 Oct 2008 22:00:29 +0000 (15:00 -0700)]
kclient: debug message

16 years agodebian: include debian dir in usr/src/modules/ceph
Sage Weil [Fri, 17 Oct 2008 01:27:57 +0000 (18:27 -0700)]
debian: include debian dir in usr/src/modules/ceph

16 years agokclient: some export.c comments
Sage Weil [Thu, 16 Oct 2008 23:39:43 +0000 (16:39 -0700)]
kclient: some export.c comments

16 years agokclient: addr.c comments, cleanup
Sage Weil [Thu, 16 Oct 2008 23:34:44 +0000 (16:34 -0700)]
kclient: addr.c comments, cleanup

I'm leaving the ability to redirty a page within a given
snapc in ceph_set_page_dirty (even though it isn't used)
because I'm not sure about the failure paths.. I suspect
we will need it later.

16 years agodebian: make ceph-source include module source
Sage Weil [Thu, 16 Oct 2008 22:33:54 +0000 (15:33 -0700)]
debian: make ceph-source include module source

There's a bunch of weirdness in the rules file, though, that I don't understand.
What's rules.modules.in?  Why was it being populated (along with a debian dir)
in /usr/src/modules/ceph?

16 years agokclient: mdsmap cleanups
Sage Weil [Thu, 16 Oct 2008 22:33:10 +0000 (15:33 -0700)]
kclient: mdsmap cleanups

16 years agokclient: osdmap.c cleanups
Sage Weil [Thu, 16 Oct 2008 22:26:24 +0000 (15:26 -0700)]
kclient: osdmap.c cleanups

16 years agokclient: osd_client comments, cleanup
Sage Weil [Thu, 16 Oct 2008 22:01:35 +0000 (15:01 -0700)]
kclient: osd_client comments, cleanup

16 years agodebian: fix kernel module in source package
Sage Weil [Thu, 16 Oct 2008 20:53:56 +0000 (13:53 -0700)]
debian: fix kernel module in source package

16 years agoceph_fs.h: whitespace
Sage Weil [Thu, 16 Oct 2008 20:20:41 +0000 (13:20 -0700)]
ceph_fs.h: whitespace

16 years agokclient: more mds_client comments, cleanup
Sage Weil [Thu, 16 Oct 2008 19:13:09 +0000 (12:13 -0700)]
kclient: more mds_client comments, cleanup

16 years agokclient: more endianity fixes
Yehuda Sadeh [Thu, 16 Oct 2008 18:54:25 +0000 (11:54 -0700)]
kclient: more endianity fixes

16 years agokclient: endianity handling fixes
Yehuda Sadeh [Thu, 16 Oct 2008 16:29:59 +0000 (09:29 -0700)]
kclient: endianity handling fixes

16 years agokclient: endianity fixes
Yehuda Sadeh [Wed, 15 Oct 2008 23:08:33 +0000 (16:08 -0700)]
kclient: endianity fixes

16 years agokclient: some mds_client.c comments, cleanup
Sage Weil [Thu, 16 Oct 2008 16:14:57 +0000 (09:14 -0700)]
kclient: some mds_client.c comments, cleanup

16 years agokclient: send osd, ping messages in lossless mode
Sage Weil [Thu, 16 Oct 2008 02:50:03 +0000 (19:50 -0700)]
kclient: send osd, ping messages in lossless mode

This is what they should have been doing all along.  Oops.

16 years agokclient: super.c comments, cleanup
Sage Weil [Thu, 16 Oct 2008 00:17:39 +0000 (17:17 -0700)]
kclient: super.c comments, cleanup

16 years agokclient: move inode alloc/destroy into inode.c
Sage Weil [Wed, 15 Oct 2008 23:46:18 +0000 (16:46 -0700)]
kclient: move inode alloc/destroy into inode.c

16 years agokclient: caps.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 23:33:35 +0000 (16:33 -0700)]
kclient: caps.c comments, cleanup

16 years agokclient: remove compilation warning
Yehuda Sadeh [Wed, 15 Oct 2008 22:15:09 +0000 (15:15 -0700)]
kclient: remove compilation warning

16 years agokclient: sparse code cleanup
Yehuda Sadeh [Wed, 15 Oct 2008 22:10:56 +0000 (15:10 -0700)]
kclient: sparse code cleanup

16 years agokclient: file.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 22:03:29 +0000 (15:03 -0700)]
kclient: file.c comments, cleanup

16 years agokclient: dir.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 21:46:26 +0000 (14:46 -0700)]
kclient: dir.c comments, cleanup

16 years agovstartnew.sh: -n (norestart) to avoid crun restart loop
Sage Weil [Wed, 15 Oct 2008 21:01:22 +0000 (14:01 -0700)]
vstartnew.sh: -n (norestart) to avoid crun restart loop

16 years agokclient: fix inode refcount problems in ceph_fill_trace
Sage Weil [Wed, 15 Oct 2008 20:52:33 +0000 (13:52 -0700)]
kclient: fix inode refcount problems in ceph_fill_trace

If we fail to take i_mutex, we get the inode via ceph_get_inode, which bumps
i_count.  Make sure we iput() it before jumping to update_inode label.  Note
that d_instantiate consumes the ref, but the other paths (d_alloc_anon,
d_find_alias) do not.

16 years agokclient: some more code cleanup
Yehuda Sadeh [Wed, 15 Oct 2008 20:34:24 +0000 (13:34 -0700)]
kclient: some more code cleanup

16 years agofix crc issue (when len < 4)
Yehuda Sadeh [Wed, 15 Oct 2008 20:33:38 +0000 (13:33 -0700)]
fix crc issue (when len < 4)

16 years agokclient: some sparse cleaning
Yehuda Sadeh [Wed, 15 Oct 2008 18:54:22 +0000 (11:54 -0700)]
kclient: some sparse cleaning

16 years agouse flags for mds chown requests
Sage Weil [Wed, 15 Oct 2008 19:12:40 +0000 (12:12 -0700)]
use flags for mds chown requests

This avoids colliding with uids of -1 (with the old, lame method).
Clean up client's utimes() code while we're at it to explicitly specify
a mask.

16 years agokclient: inode.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 19:02:20 +0000 (12:02 -0700)]
kclient: inode.c comments, cleanup

16 years agokclient: fix umount_begin for 2.6.26+
Yehuda Sadeh [Wed, 15 Oct 2008 18:52:20 +0000 (11:52 -0700)]
kclient: fix umount_begin for 2.6.26+

16 years agofix wrong byte ordering handling (thanks brent!)
Yehuda Sadeh [Wed, 15 Oct 2008 16:59:31 +0000 (09:59 -0700)]
fix wrong byte ordering handling (thanks brent!)

16 years agoremove debug helper code for crc32
Yehuda Sadeh [Wed, 15 Oct 2008 16:58:30 +0000 (09:58 -0700)]
remove debug helper code for crc32

16 years agokclient: preallocate cap if we expect it
Sage Weil [Tue, 14 Oct 2008 23:53:38 +0000 (16:53 -0700)]
kclient: preallocate cap if we expect it

Preallocate memory for a capability when doing an open request.  If we don't
use it, free it with the mds_request.  This avoids the possibility of an ENOMEM
in the reply handler, which is a little difficult to deal with intelligently.

16 years agokclient: remove static caps from inode
Sage Weil [Tue, 14 Oct 2008 23:18:24 +0000 (16:18 -0700)]
kclient: remove static caps from inode

Since most inodes will never have a capability, there is no point in embedding them
in the ceph_inode_info and paying the memory overhead in the general case.

16 years agokclient: super.h comments, cleanup, some minor cruft removed
Sage Weil [Tue, 14 Oct 2008 23:15:54 +0000 (16:15 -0700)]
kclient: super.h comments, cleanup, some minor cruft removed

16 years agoceph_fs: some cleanup, including mon statfs messages
Sage Weil [Tue, 14 Oct 2008 22:17:14 +0000 (15:17 -0700)]
ceph_fs: some cleanup, including mon statfs messages

16 years agochecking buffer before calculating the crc (for debug)
Yehuda Sadeh [Tue, 14 Oct 2008 21:53:56 +0000 (14:53 -0700)]
checking buffer before calculating the crc (for debug)

16 years agokclient: some more error handling
Yehuda Sadeh [Tue, 14 Oct 2008 21:49:58 +0000 (14:49 -0700)]
kclient: some more error handling

16 years agokclient: remove ping_ack entirely
Sage Weil [Tue, 14 Oct 2008 19:37:32 +0000 (12:37 -0700)]
kclient: remove ping_ack entirely

16 years agodstart.sh: use unstable btrfs
Sage Weil [Tue, 14 Oct 2008 20:43:51 +0000 (13:43 -0700)]
dstart.sh: use unstable btrfs

16 years agofilestore: fix btrfs version detection
Sage Weil [Tue, 14 Oct 2008 20:05:30 +0000 (13:05 -0700)]
filestore: fix btrfs version detection

16 years agokclient: lower debug level on debug message
Yehuda Sadeh [Tue, 14 Oct 2008 19:48:52 +0000 (12:48 -0700)]
kclient: lower debug level on debug message

16 years agokclient: fix race condition
Yehuda Sadeh [Tue, 14 Oct 2008 19:34:18 +0000 (12:34 -0700)]
kclient: fix race condition

16 years agokclient: remove ping ack
Yehuda Sadeh [Tue, 14 Oct 2008 16:20:25 +0000 (09:20 -0700)]
kclient: remove ping ack

16 years agokclient: preliminary forced umount
Yehuda Sadeh [Tue, 14 Oct 2008 01:37:39 +0000 (18:37 -0700)]
kclient: preliminary forced umount

16 years agokclient: ping osds whose requests are timing out
Sage Weil [Tue, 14 Oct 2008 00:42:44 +0000 (17:42 -0700)]
kclient: ping osds whose requests are timing out

This ensures the client detects when the osd tcp session has closed and
a reply may have been lost.  If that does happen, the handle_reset handler
will be called and the request(s) will be resent.

16 years agomsgr: ping needs no payload. lose ping_ack.
Sage Weil [Tue, 14 Oct 2008 00:41:07 +0000 (17:41 -0700)]
msgr: ping needs no payload.  lose ping_ack.

They were unused/ignored.

16 years agokclient: convert spinlock to mutex in osdc
Sage Weil [Mon, 13 Oct 2008 23:57:44 +0000 (16:57 -0700)]
kclient: convert spinlock to mutex in osdc

16 years agokclient: resend osd requests if osd resets the connection
Sage Weil [Mon, 13 Oct 2008 23:19:35 +0000 (16:19 -0700)]
kclient: resend osd requests if osd resets the connection

16 years agokclient: some error handling
Yehuda Sadeh [Mon, 13 Oct 2008 23:09:39 +0000 (16:09 -0700)]
kclient: some error handling

16 years agoassert: include rudimentary stack dump
Sage Weil [Mon, 13 Oct 2008 20:49:15 +0000 (13:49 -0700)]
assert: include rudimentary stack dump

16 years agocustom assertion handler
Sage Weil [Mon, 13 Oct 2008 20:32:03 +0000 (13:32 -0700)]
custom assertion handler

Print assertion error to log file, if it is open.

16 years agoobjecter: ping laggy osds
Sage Weil [Mon, 13 Oct 2008 19:53:28 +0000 (12:53 -0700)]
objecter: ping laggy osds

This ensures we detect any session resets.  This necessary because the
osd message policy on replies is lossy.

16 years agoosd: lossy send policy on acks
Sage Weil [Mon, 13 Oct 2008 19:51:42 +0000 (12:51 -0700)]
osd: lossy send policy on acks

16 years agoobjecter: kick requests when we detect a remote session reset
Sage Weil [Mon, 13 Oct 2008 19:51:11 +0000 (12:51 -0700)]
objecter: kick requests when we detect a remote session reset

16 years agomsgr: handle incoming lossy connection when outgoing policy is lossless
Sage Weil [Mon, 13 Oct 2008 19:50:06 +0000 (12:50 -0700)]
msgr: handle incoming lossy connection when outgoing policy is lossless

Basically, try to wake up our lossless connection.  The lossy attempt never
survives.  This avoids worrying about keeping session state alive (connect_seq,
etc.).

16 years agomsgr: track lossy mode independently for self, peer
Sage Weil [Mon, 13 Oct 2008 18:56:31 +0000 (11:56 -0700)]
msgr: track lossy mode independently for self, peer

The policy will be asymmetrical for the OSDs, so we need to
track it independently.

The various assertions aren't all worked out yet.  Notably,
connection races aren't quite right.  But the basic bits
are there.

16 years agoclient: fix && || parens
Sage Weil [Thu, 9 Oct 2008 23:22:23 +0000 (16:22 -0700)]
client: fix && || parens

16 years agofilestore: use BTRFS_IOC_SYNC instead of fsync
Sage Weil [Thu, 9 Oct 2008 23:03:57 +0000 (16:03 -0700)]
filestore: use BTRFS_IOC_SYNC instead of fsync

fsync() won't do the right thing on newer versions of btrfs, due do the
new tree log stuff.  Use the SYNC ioctl.

16 years agofilestore: detect new btrfs ioctls
Sage Weil [Thu, 9 Oct 2008 23:02:50 +0000 (16:02 -0700)]
filestore: detect new btrfs ioctls

Don't use them yet, though.

16 years agofix of crc32c calculation
Yehuda Sadeh [Mon, 13 Oct 2008 18:45:14 +0000 (11:45 -0700)]
fix of crc32c calculation

16 years agoreplace the crc32c code
Yehuda Sadeh [Fri, 10 Oct 2008 22:52:30 +0000 (15:52 -0700)]
replace the crc32c code

16 years agodebian: changelog update
Sage Weil [Thu, 9 Oct 2008 20:59:02 +0000 (13:59 -0700)]
debian: changelog update

16 years agodebian: fix debian-doc; include debian dir in dist tarball
Sage Weil [Thu, 9 Oct 2008 20:34:32 +0000 (13:34 -0700)]
debian: fix debian-doc; include debian dir in dist tarball

16 years agomon: populate last_beacon properly for mds liveness check
Sage Weil [Thu, 9 Oct 2008 18:58:10 +0000 (11:58 -0700)]
mon: populate last_beacon properly for mds liveness check

It was only populating on paxos updates, but that missed the case where the
leader moves from one monitor to another (live) peon, who has an empty last_beacon.
Move it where it makes the most sense.

16 years agodebian: tweak ceph-doc paths
Sage Weil [Thu, 9 Oct 2008 18:27:07 +0000 (11:27 -0700)]
debian: tweak ceph-doc paths

16 years agodstart.sh: replace dstartnew.sh
Sage Weil [Thu, 9 Oct 2008 05:06:24 +0000 (22:06 -0700)]
dstart.sh: replace dstartnew.sh

-n will create new fs, otherwise it will start up using existing data.

16 years agofilestore: fix collection attrs
Sage Weil [Thu, 9 Oct 2008 05:05:35 +0000 (22:05 -0700)]
filestore: fix collection attrs

Need to prepend user.ceph. to attr names...

16 years agotodos
Sage Weil [Wed, 8 Oct 2008 23:45:31 +0000 (16:45 -0700)]
todos

16 years agomon: include pg usage stats, header in pg dump
Sage Weil [Wed, 8 Oct 2008 23:31:50 +0000 (16:31 -0700)]
mon: include pg usage stats, header in pg dump

16 years agodebian: move map tools and crun to ceph-common
Sage Weil [Wed, 8 Oct 2008 23:18:10 +0000 (16:18 -0700)]
debian: move map tools and crun to ceph-common

16 years agoosd: account for used kb separately
Sage Weil [Wed, 8 Oct 2008 19:10:16 +0000 (12:10 -0700)]
osd: account for used kb separately

Account for used separately from available, as the underlying fs may do tricks
reserving space for root and such.

16 years agoosd: use kb instead of blocks for pg, osd utilization
Sage Weil [Wed, 8 Oct 2008 18:43:08 +0000 (11:43 -0700)]
osd: use kb instead of blocks for pg, osd utilization

16 years agoosd: prioritize heartbeat messages
Sage Weil [Wed, 8 Oct 2008 17:52:46 +0000 (10:52 -0700)]
osd: prioritize heartbeat messages

16 years agomsgr: include priority in msg header, make dispatch a priority queue
Sage Weil [Wed, 8 Oct 2008 17:49:12 +0000 (10:49 -0700)]
msgr: include priority in msg header, make dispatch a priority queue

Generalizes previous hack that put messages from the monitor at the front of the
dispatch queue.  Monitor now just sets a (non-default) default_send_priority
of CEPH_MSG_PRIO_HIGH.  That value is used only if the message priority isn't
set explicitly by set_priority() before being queued for send.

16 years agojournal: protect journal access (namely, completions) with mutex
Sage Weil [Wed, 8 Oct 2008 16:47:37 +0000 (09:47 -0700)]
journal: protect journal access (namely, completions) with mutex

The commit_waiters map was getting corrupted occasionally.

16 years agodebian: include crun in osd, mds, mon packages
Sage Weil [Wed, 8 Oct 2008 16:31:23 +0000 (09:31 -0700)]
debian: include crun in osd, mds, mon packages

16 years agoosdmaptool: fix silly num_dom bug
Sage Weil [Wed, 8 Oct 2008 16:28:58 +0000 (09:28 -0700)]
osdmaptool: fix silly num_dom bug

16 years agoceph.spec.in: include crun
Erwin, Brock A [Wed, 8 Oct 2008 16:28:32 +0000 (09:28 -0700)]
ceph.spec.in: include crun

16 years agokclient: initializing kaddr (merge fix) v0.4
Yehuda Sadeh [Tue, 7 Oct 2008 20:18:52 +0000 (13:18 -0700)]
kclient: initializing kaddr (merge fix)