Validated the label and verified the existence of the label at the beginning of build_branch().
This will avoid:
* Assigning invalid testing labels like 'abcd' and other existing labels like 'bug', 'cleanup' etc; as testing labels
* Stops spamming the Label list in the Ceph repo by creating new (invalid format) labels everytime. The labels can be created manually in github and ptl-tool only uses them.
Signed-off-by: Jos Collin <jcollin@redhat.com>
We were passing this error back to rocksdb, which would then crash (much)
later with a corrupted SST file! No good.
Signed-off-by: Sage Weil <sage@redhat.com>
I'm not quite sure why the FunctionContext did not ever execute on the
finisher thread (perhaps the [&] captured some state on the stack that it
shouldn't have?). In any case, using a traditional Context here appears
to resolve the problem (of the async deferred_try_submit() never executing,
leading to a bluestore stall/deadlock).
Fixes: http://tracker.ceph.com/issues/21470
Signed-off-by: Sage Weil <sage@redhat.com>
We need to wake the kv thread whenever setting deferred_aggressive to
ensure that txns with deferred io that have committed but haven't submitted
their deferred writes get submitted. This aligns us with the other
users of deferred_aggressive (e.g., _osr_drain_all).
Signed-off-by: Sage Weil <sage@redhat.com>
Right now we transparently map a RemoteReservationRejected into a
*Canceled event because this what peers send over the wire. Even
once new peers start sending and explicit CANCEL, old peers will
still do so, so we'll maintain this mapping for a while.
Signed-off-by: Sage Weil <sage@redhat.com>
This reduces pg->reject_reservation() callsites from 2 to 1 and
makes the state transitions a bit more explicit.
Signed-off-by: Sage Weil <sage@redhat.com>
when ./do_cmake.sh add -DWITH_DPDK=ON -DUSE_CRYPTOPP=ON -DWITH_NSS=OFF -DWITH_TESTS=OFF
if build debian package, need remove some file froms install since we disabled TESTS.
Signed-off-by: chunmei <chunmei.liu@intel.com>
The snap_lock is being held when this method is invoked, which can
result in a deadlock.
Fixes: http://tracker.ceph.com/issues/21670
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
This is fairly clean: we define an RAII object in the Messenger.h on
Solaris, and "declare" it with a macro in the implementations. There's
no code duplication and on Linux it's just entirely compiled out.
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
This reverts commit 131deb3976. It added
code that shouldn't have entered the project repo.
Conflicts:
src/msg/async/AsyncConnection.cc
src/msg/async/AsyncConnection.h
src/msg/simple/Pipe.cc
src/msg/simple/Pipe.h
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
The 'osd pool create' arg parsing is broken; the rule name for
'ceph osd pool create $name $numpgs replicated $rulename' is passed
via the erasure_code_profile param. Too many req=false options
without a way to disambiguate them.
Work around it by passing both 'rule' and 'erasure_code_profile'
keys, so that if/when the hack in OSDMonitor.cc is removed it will
still work. Blech.
Signed-off-by: Sage Weil <sage@redhat.com>
With -y you can't see the entire line when it is long, which is
needed to identify the diff failure in
http://tracker.ceph.com/issues/21618
Instead, let the interactive user specify the option if they want it.
Signed-off-by: Sage Weil <sage@redhat.com>
