]>
git.apps.os.sepia.ceph.com Git - ceph.git/log
Sage Weil [Tue, 12 Jun 2012 19:48:37 +0000 (12:48 -0700)]
Merge branch 'wip-keyrings'
Reviewed-by: Greg Farnum <greg@inktank.com>
Sage Weil [Fri, 8 Jun 2012 19:45:42 +0000 (12:45 -0700)]
ceph_argparse: make entityname parsing error more helpful
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 02:42:51 +0000 (19:42 -0700)]
rgw: default keyring to $rgw_data/keyring
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 02:50:50 +0000 (19:50 -0700)]
vstart.sh: set data dirs instead of keyring locations
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 1 Jun 2012 18:02:09 +0000 (11:02 -0700)]
new default keyring location for daemons
$osd_data/keyring
$mds_data/keyring
Fixes: #2498
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 1 Jun 2012 17:58:23 +0000 (10:58 -0700)]
mds: add mds_data config
This is where the keyring will live. Eventually more may move in too.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sun, 10 Jun 2012 02:40:52 +0000 (19:40 -0700)]
config: improve variable substitution
- allow ${foo_bar}
- prevent $foogarbage from substitution variable 'foo'
- use std::string throughout
- improve tests
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 19:45:30 +0000 (12:45 -0700)]
config: expand any config variable in config
This is currently broken wrt 'foo' vs 'foo_bar', but otherwise works.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 1 Jun 2012 19:44:42 +0000 (12:44 -0700)]
config: make meta expansion in set_val optional
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 26 May 2012 05:30:12 +0000 (22:30 -0700)]
osdcaps: drop dead peer_type code
Not sure what this wasn intended for.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 02:30:19 +0000 (19:30 -0700)]
osd: simplify session peer type check
Check the authenticated peer name directly; don't store a copy in the
OSDCaps struct. Also set the entity_name member of the session!
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 26 May 2012 05:36:49 +0000 (22:36 -0700)]
entity_name: is_*() methods
Signed-off-by: Sage Weil <sage@inktank.com>
caleb miles [Tue, 12 Jun 2012 18:15:57 +0000 (11:15 -0700)]
crush: fix retry histogram array overflow due to indexing error
Signed-off-by: caleb miles <caleb.miles@inktank.com>
Sage Weil [Tue, 12 Jun 2012 18:28:04 +0000 (11:28 -0700)]
cls_rgw: use updated internal omap api
These were changed by
23ebc093a64a9c9812124ba6b87e301f95a53a5c .
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 12 Jun 2012 18:27:36 +0000 (11:27 -0700)]
cls_rgw: fix CLS_LOG macro usage
Now takes a level.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 11 Jun 2012 23:25:42 +0000 (16:25 -0700)]
remove ceph-kdump-copy
Moved to its own repo.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 11 Jun 2012 22:53:36 +0000 (15:53 -0700)]
debian: limit build archs
We depend on leveldb, which only builds on
amd64 armel armhf i386 ia64 mipsel
Reported-by: Laszlo Boszormenyi (GCS) <gcs@debian.hu>
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 12 Jun 2012 01:36:59 +0000 (18:36 -0700)]
Merge remote-tracking branch 'gh/wip-2516-2'
Reviewed-by: Sage Weil <sage@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 23:39:15 +0000 (16:39 -0700)]
rgw: usage statistics also count num of ops/successful_ops
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Laszlo Boszormenyi (GCS) [Mon, 11 Jun 2012 22:48:50 +0000 (15:48 -0700)]
debian: move ./configure to correct rules section
Signed-off-by: Laszlo Boszormenyi (GCS) <gcs@debian.hu>
Sage Weil [Mon, 11 Jun 2012 22:48:38 +0000 (15:48 -0700)]
debian: remove dup python build-dep
Reported-by: Laszlo Boszormenyi (GCS) <gcs@debian.hu>
Signed-off-by: Sage Weil <sage@inktank.com>
Tommi Virtanen [Mon, 11 Jun 2012 22:27:02 +0000 (15:27 -0700)]
upstart: Read crush location and weight from ceph.conf.
This introduces two new config variables, osd_crush_location
and osd_crush_weight. Not currently included in config_opts.h,
as these are not used in the C++ code.
Signed-off-by: Tommi Virtanen <tv@inktank.com>
Sage Weil [Mon, 11 Jun 2012 22:21:11 +0000 (15:21 -0700)]
mon: allow keys to be cleared in ceph osd crush set ... loc key/values
If we encounter a 'key=' with no value, clear any previous value for that
key, so that
'a=foo b=bar a='
is equivalent to
'b=bar'
Fixes: #2540
Signed-off-by: Sage Weil <sage@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 22:13:18 +0000 (15:13 -0700)]
rgw: fix cli test
Get it up to date with latest cli changes.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 22:09:00 +0000 (15:09 -0700)]
rgw: usage log cleanups
Also, recalc_round_timestamp should be done under lock.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 21:59:38 +0000 (14:59 -0700)]
Merge pull request #13 from fghaas/rgw-manpage
doc: update examples in radosgw-admin man page
Florian Haas [Mon, 11 Jun 2012 21:49:11 +0000 (23:49 +0200)]
doc: update examples in radosgw-admin man page
* "user gen" is gone and is now "user create"
* --uid option is mandatory for "user create"
Mailing list thread: http://www.spinics.net/lists/ceph-devel/msg06744.html
Signed-off-by: Florian Haas <florian@hastexo.com>
Yehuda Sadeh [Mon, 11 Jun 2012 21:25:01 +0000 (14:25 -0700)]
rgw: rgw_rados related fixes for usage logging
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 21:00:20 +0000 (14:00 -0700)]
rgw/rgw_cls_api.h cleanup
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 18:03:57 +0000 (11:03 -0700)]
rgw_admin: usage admin commands
- radosgw-admin usage show
- radosgw-admin usage trim
Both commands can get --start-date, --end-date, --uid
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 18:01:56 +0000 (11:01 -0700)]
rgw: usage logger
Accumulate usage info and flush it periodically.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 17:23:49 +0000 (10:23 -0700)]
rgw: access methods for new usage ops
Wrappers for new rados class methods, and new RGWRados
methods to handle usage functionality.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 17:14:43 +0000 (10:14 -0700)]
rgw: new config options
New config options for usage logging:
- rgw_enable_usage_log: enable usage logging
- rgw_usage_log_flush_threshold - limit on number of pending updates
before synchronously flushing update
- rgw_usage_log_tick_interval - asynchronous flush interval
- rgw_usage_max_shards - split info across that many objects
- rgw_usage_max_user_shards - split single user info across that many
objects
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Mon, 11 Jun 2012 17:11:17 +0000 (10:11 -0700)]
rgw: new class methods for handling usage information
The new methods are:
- user_usage_log_add: add new usage information
- user_usage_log_read: get usage information
- user_usage_log_trim: remove usage information
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Wed, 6 Jun 2012 21:42:12 +0000 (14:42 -0700)]
rgw: don't fail initialization if socket path returns ENXIO
ENXIO is expected when trying to read the unix domain socket
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sage Weil [Mon, 11 Jun 2012 17:36:01 +0000 (10:36 -0700)]
Merge remote-tracking branch 'gh/wip-crush-tunables'
Josh Durgin [Mon, 11 Jun 2012 06:21:58 +0000 (23:21 -0700)]
cls_rbd: add get_all_features method
This is useful for reporting which features an osd supports, and for
testing rados_exec. Update the rados api tests to use this method
instead of test_exec, which was removed.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>
Yan, Zheng [Mon, 11 Jun 2012 02:33:00 +0000 (10:33 +0800)]
mon: fix pg state logging
PGMap->num_pg_by_state is a PG state to number of PG in the state
mapping. PGMonitor::update_logger wrongly interprets the mapping.
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Sage Weil <sage@inktank.com>
Josh Durgin [Mon, 11 Jun 2012 05:01:19 +0000 (22:01 -0700)]
workunits/rbd: add workunit for running cls_rbd tests
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Mon, 11 Jun 2012 04:59:07 +0000 (21:59 -0700)]
workunits/rbd: disable remove_with_watcher test
This will fail until #2533 is fixed.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Sun, 10 Jun 2012 21:12:16 +0000 (14:12 -0700)]
Merge remote-tracking branch 'gh/wip-rbd-format'
Josh Durgin [Sun, 10 Jun 2012 00:27:31 +0000 (17:27 -0700)]
librbd: remove unnecessary notify from add_snap()
The only caller, snapshot_add(), already does a notify when add_snap()
succeeds.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 10 Jun 2012 00:16:45 +0000 (17:16 -0700)]
librbd: ignore RBD_MAX_BLOCK_NAME_SIZE when generating object ids
The actual data object ids don't need to be artificially restricted in
length. RBD_MAX_BLOCK_NAME_SIZE just limits the size of the object
prefix, since it's used in rbd_info_t.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:43:43 +0000 (08:43 -0700)]
workunits/rbd: allow creating images in different formats
This will allow adding more arguments for testing combinations
of features in the future.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:40:27 +0000 (08:40 -0700)]
rados: add commands to interact with object maps
The input values are stored as-is, and any values read are dumped in
hex. Rename listomap to listomapkeys to distinguish from
listomapvalues. Also add it to the man page.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:31:14 +0000 (08:31 -0700)]
test_librbd, test_rbd.py: optionally use the new format
Read the desired features from an environment variable, RBD_FEATURES.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:09:35 +0000 (08:09 -0700)]
rbd.py: add support for creating images in the new format
The new arguments are optional, so they are backwards compatible.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:07:40 +0000 (08:07 -0700)]
rbd: update for the new format
No features exist right now, so there are no extra options for them.
The old format is still used by default, and since the default will
change with layering, --new-format will be removed at that point and is
intentionally left undocumented.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:43:32 +0000 (07:43 -0700)]
librbd: add create2 to create an image with the new format
This will fail if features are requested that the client or server
does not support. Currently there are no features defined, so
zero is the only valid value.
copy() preserves the format and features of the source image.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
John Wilkins [Sat, 9 Jun 2012 21:45:05 +0000 (14:45 -0700)]
doc: Segregated set up git from clone.
Signed-off-by: John Wilkins <john.wilkins@inktank.com>
Sage Weil [Sat, 9 Jun 2012 05:10:43 +0000 (22:10 -0700)]
debian: ceph depends on python
Due to the new ceph-disk-* scripts. Newer dists are smart enough to figure
this out, but maverick is not.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 9 Jun 2012 04:37:02 +0000 (21:37 -0700)]
CrushTester: drop range marking
This is subsumed by mark-down-ratio 1 with mark-down-bucket-ratio N.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 9 Jun 2012 05:29:02 +0000 (22:29 -0700)]
crushtool: drop useless clitest
This is an ancient test for an old 'bug' in functionality we're removing.
Also, it is sensitive to tester output, which will be changing a lot in
the coming weeks/months.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 9 Jun 2012 03:39:41 +0000 (20:39 -0700)]
CrushTester: simplify, clean up mark down
- put it in a separate function
- operate on temporary weight vector, not user-modified input
- guard the whole thing with an #ifdef
- permute candidates and use first N, to ensure we end up picking the right
number of buckets/items.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 9 Jun 2012 03:25:47 +0000 (20:25 -0700)]
CrushTester: some whitespace
Signed-off-by: Sage Weil <sage@inktank.com>
caleb miles [Sat, 9 Jun 2012 03:20:05 +0000 (20:20 -0700)]
CrushTester: randomize mark_down behavior
When marking a batch of devices down, select which buckets to visit sort
of randomly.
Signed-off-by: caleb miles <caleb.miles@inktank.com>
caleb miles [Sat, 9 Jun 2012 03:19:36 +0000 (20:19 -0700)]
CrushTester: remove dead code
Signed-off-by: caleb miles <caleb.miles@inktank.com>
caleb miles [Sat, 9 Jun 2012 03:19:18 +0000 (20:19 -0700)]
CrushTester: some cleanup
Signed-off-by: caleb miles <caleb.miles@inktank.com>
caleb miles [Sat, 9 Jun 2012 03:17:52 +0000 (20:17 -0700)]
CrushTester: mark_down_ratio and mark_down_bucket_ratio
Move functionality to allow user to control bucket vs device mark-down
probabilities independently.
Signed-off-by: caleb miles <caleb.miles@inktank.com>
Sage Weil [Wed, 23 May 2012 01:54:10 +0000 (18:54 -0700)]
mon: add 'auth get-or-create' variant that outputs a keyring
This is more convenient when you're creating daemons, since you want to
write out a keyring file for them. This lets you do it in a single
command.
Signed-off-by: Sage Weil <sage@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:56:59 +0000 (07:56 -0700)]
librbd: use ImageCtx members instead of the old header in resize()
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:55:52 +0000 (07:55 -0700)]
librbd: validate order before creating an image
The value must be passed, and it shouldn't be below 4k
(enforced by the command line tool already) or above the
range expressible in the header.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:54:02 +0000 (07:54 -0700)]
librbd: rename md_oid parameters to header_oid
This is more consistent with the rest of the code now,
and is a bit more clear.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:53:07 +0000 (07:53 -0700)]
librbd: make rename work with any header format
Instead of interpreting the header, just copy all the data and
omap values from the original header to the newly name one.
This will continue working with future header changes.
We can create the new header and write all data and omap values
to it atomically to avoid some races.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 14:49:04 +0000 (07:49 -0700)]
librbd: use cls_client functions for calling class methods
Use the old or new methods make resize, snapshot add and snapsnhot
remove work with both old and new formats.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 07:05:41 +0000 (00:05 -0700)]
librbd: remove on-disk header argument from helper functions
Make most of them take the parameters they actually use.
trim_image() now takes an ImageCtx, which means remove() must
open the image. This has the nice side effect of not duplicating
the snapshot listing code for the old format.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:04:12 +0000 (08:04 -0700)]
librbd: check that the current snapid for a snap name matches
Checking that it exists doesn't prevent you from having the snapshot
change out from under you in the following situation:
You have the image open at snapshot "foo".
Someone removes snapshot "foo", writes some data to the image, and
creates a new snapshot called "foo".
This second snapshot will have a different id, but nothing prevents it
from having the name of a previously deleted snapshot.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 15:01:19 +0000 (08:01 -0700)]
librbd: update ictx_refresh to work with both formats
It now sets the member variables of ImageCtx so other functions
don't have to use the on-disk header. If the features use by
the new format are incompatible with this client, an error is returned.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 06:14:17 +0000 (23:14 -0700)]
librbd: Update ImageCtx for new format
Detect the format when an image is opened by the presence of the
original format header object. Use member variables of ImageCtx to
store image metadata instead of the on-disk header format
ImageCtx::header.
This lays the foundation for changing the rest of librbd to work with
old and new formats.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 8 Jun 2012 05:54:45 +0000 (22:54 -0700)]
cls_rbd: add methods for interacting with the new header format
Add the client side (cls_rbd_client) for testing as well.
librbd will use the functions in cls_rbd_client to interact with cls_rbd.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Mon, 4 Jun 2012 02:42:23 +0000 (19:42 -0700)]
librbd: remove useless ENOMEM checks
There will be an exception if memory can't be allocated.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:45:15 +0000 (03:45 -0700)]
DBObjectMap: remove extra semicolon
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:43:40 +0000 (03:43 -0700)]
vstart: debug osd classes when debugging is on
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:42:43 +0000 (03:42 -0700)]
objclass: add methods to get keys without values, and set multiple values
These will be useful in the new rbd header format. This brings the
API into parity with the available OSD ops for omaps.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:41:02 +0000 (03:41 -0700)]
objclass: rename omap methods to be consistent with other interfaces
The osd ops and the librados methods all distinguish between keys
and values. It's useful to be able to read a bunch of keys without
fetching the values sometimes too.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:35:38 +0000 (03:35 -0700)]
objclass: allow classes to log at different levels
Add a logging level parameter to CLS_LOG, and a CLS_ERR that's
analagous to derr.
For now cls_rgw is all logging at level 1, preserving existing
behavior. I'm more familiar with cls_rbd, so I changed those
logging levels.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:27:32 +0000 (03:27 -0700)]
class_debug: add indent settings header
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:26:47 +0000 (03:26 -0700)]
test_librbd: don't include a .cc file
Add it as a source like the rest of the gtests.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Sun, 3 Jun 2012 10:25:13 +0000 (03:25 -0700)]
cls_rgw: add indent settings header
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Wed, 30 May 2012 03:13:52 +0000 (20:13 -0700)]
objclass: pass strings as const references
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Wed, 30 May 2012 00:33:58 +0000 (17:33 -0700)]
objclass: add indent settings header
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 25 May 2012 21:12:18 +0000 (14:12 -0700)]
cls_rbd: add indent settings header
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 25 May 2012 21:09:02 +0000 (14:09 -0700)]
cls_rbd: remove unused test_exec and snap_revert methods
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Fri, 8 Jun 2012 19:48:02 +0000 (12:48 -0700)]
mon: require force argument for 'mds newfs ...'
This is dangerous, make sure they know what they're doing. I bet <someone>
$1 that Greg asked me to do this originally.
Signed-off-by: Sage Weil <sage@inktank.com>
caleb miles [Fri, 8 Jun 2012 18:34:38 +0000 (11:34 -0700)]
Cleaned some lingering references to "verbose"
Signed-off-by: caleb miles <caleb.miles@inktank.com>
Sage Weil [Fri, 8 Jun 2012 18:18:59 +0000 (11:18 -0700)]
Merge branch 'wip-crush-tunables'
Reviewed-by: Sam Just <sam.just@inktank.com>
Sage Weil [Fri, 8 Jun 2012 18:17:14 +0000 (11:17 -0700)]
crushtool: refine tunables warning message
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 18:02:44 +0000 (11:02 -0700)]
re-include assert after boost::pool
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 17:56:57 +0000 (10:56 -0700)]
crush: compile/decompile tunables
Lines like
tunable <name> <value>
will adjust tunables. Right now only three are defined.
--enable-unsafe-tunables is required.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 17:57:28 +0000 (10:57 -0700)]
crushtool: add --enable-unsafe-tunables option
This is required to adjust tunables.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 15:52:43 +0000 (08:52 -0700)]
crushtool: --show-* instead of --output-*
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 02:33:14 +0000 (19:33 -0700)]
CrushTester: clean up output interface
Multiple accessors. Init in ctor. Avoid temp vars in crushtool.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 02:21:51 +0000 (19:21 -0700)]
CrushTester: dump histogram of choose attempts
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 8 Jun 2012 02:21:36 +0000 (19:21 -0700)]
crush: generate histogram of choose tries
Optionally populate a histogram of choose descent attempts.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 23:52:57 +0000 (16:52 -0700)]
crush: fix leaf recursion if we already collided
This just saves us some cycles, but does not effect placement results at
all.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 23:34:11 +0000 (16:34 -0700)]
CrushTester: optionally output bad mappings
Optionally dump bad inputs to stdout.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 23:08:23 +0000 (16:08 -0700)]
crushtool: arguments to adjust tunables
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 22:57:09 +0000 (15:57 -0700)]
crush: make magic numbers tunable
We have three magic numbers in crush_choose that are now tunable. The
first two control the local retry behavior, including fallback to a
permutation. The last is the total map descent attempts.
We can avoid a drastic incompatibility by making these tunable and encoded
in the map. That means users can enable/disable local retry, for example,
without changing the code. As long as the clients understand the tunables,
they can be adjusted.
This patch doesn't address the compatibility and feature bit issue. We may
want to roll that into a larger revision with more drastic changes, once
we know what those changes will look like. However, a careful user can
use the new code and modify the behavior.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 20:57:10 +0000 (13:57 -0700)]
mon: use mode 0600 throughout
Fixes: #2526
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 7 Jun 2012 19:22:47 +0000 (12:22 -0700)]
Merge remote-tracking branch 'gh/mon-auth'
Reviewed-by: Greg Farnum <greg@inktank.com>