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>
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>
Nothing has used the EboFS object storage in years.
Fix usage string of dupstore.
FileJournal and Journal are now part of OSD proper, not
EboFS.
The old scripts in src/jobs are probably broken, but
not touching them further right now.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
Fix up other places that failed to include atomic.h even
if they used atomic_t directly.
Fixes: #979
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
This reverts commit d467fbfd17.
The linkmerge stuff _is_ needed for reintegrating multi-linked stray
inodes back into the hierarchy. E.g.,
touch a
ln a b
rm a
The MDS will want to renamed the stray inode to the b dentry.
Conflicts:
src/mds/Server.cc
Signed-off-by: Sage Weil <sage@newdream.net>
md_config_t: store a ConfFile by value rather than by pointer.
ConfFile::parse_file: negate error when returning errors.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
A new parser for ConfUtils that outputs a list of syntax errors when
files are wrong.
It also checks that configuration files are valid utf-8 and don't contain
embedded nulls.
Also added a few more unit tests.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
We don't actually need to stat the object (again) since we have the size
(and mtime) in the object_info_t. Keep the existing stat in there, tho,
and assert the answer matches for the time being, since I saw behavior
that suggested they were out of sync. See #933.
Signed-off-by: Sage Weil <sage@newdream.net>
There are important cases where the replica will send a missing set and
empty log to the primary during peer (e.g., when the primary asks for it).
In that case, we need to go into pg->proc_replica_log so that peer_missing
gets updated and peering can complete. It is okay from proc_replica_log's
perspective to pass in an empty log; it will have no effect. Fix the
if() guard appropriately.
Note that the only path into _process_pg_info where missing is NULL is
the handle_pg_info path, which is used for primary->replica "activate now"
messages, updates after already active, and for replica->primary "ok i
activated" messages.
Signed-off-by: Sage Weil <sage@newdream.net>
md_config_t::get_val_from_config_file now has an option that causes it
to expand metavariables. This is necessary because many callers want
metavariables to be expanded in the results they get back.
md_config_t::set_val should expand metavariables.
Remove some unecessary typecasts from
md_config_t::get_val_from_config_file()
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
Change the ConfUtils interface so that we have a way to return
information about parse errors.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
common_init: Do metavariable expansion as a separate step after the
configuration, argv, and env have been read. This ensures that we get
the correct values for the metavariables.
Change librados to follow suit.
config: Enforce the "no metavariables in default values" restriction.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
