- make sure to always clean up excess meta replicas not just when their
number get decreased
- make sure NotServingRegionException is handled properly even when
wrapped
- add test
Signed-off-by: Peter Somogyi <psomogyi@apache.org>
Revert "HBASE-24648 Remove the legacy 'forceSplit' related code at region server side (#1990)"
This reverts commit bfec964aef09a6474d35b379fb4000cf3e8f86cc.
* HBASE-20819 Use TableDescriptor to replace HTableDescriptor in hbase-shell module
* Resolve whitespace issues
* Use removeValue rather than setValue where possible in hbase-shell
* Add String removeValue overload to TableDescriptorBuilder
* Convert remaining usages of HTableDescriptor in security.rb
* Rename hcd (HColumnDescriptor) to cfd (ColumnFamilyDescriptor)
* Add back 3 constants to HBaseConstants in hbase-shell
* Fix bad constant reference
Signed-off-by: Sean Busbey <busbey@apache.org>
Signed-off-by: stack <stack@apache.org>
Signed-off-by: Duo Zhang <zhangduo@apache.org>
* HBASE-24382 Flush partial stores of region filtered by seqId when archive wal due to too many wals
* fix checkstyle and javadoc issue
* fix javadoc issues
* move the geting of stores to HRegion, since it should not be part of FlushPolicy, and comment fix
* fix checkstyle issue
* add some comment
* remove the forceFlushAllStores since we can use families to determine how to select stores to flush
Signed-off-by: Duo Zhang <zhangduo@apache.org>
Signed-off-by: stack <stack@duboce.net>
Writing a test for this is tricky. There is enough coverage for
functional tests. Only concern is performance, but there is enough
logging for it to detect timed out/badly performing sync calls.
Additionally, this patch decouples the ZK event processing into it's
own thread rather than doing it in the EventThread's context. That
avoids deadlocks and stalls of the event thread.
Signed-off-by: Andrew Purtell <apurtell@apache.org>
Signed-off-by: Viraj Jasani <vjasani@apache.org>
Purge query Master for table descriptors; make do w/ generic options.
Logging cleanup.
hbase-server/src/main/java/org/apache/hadoop/hbase/wal/BoundedRecoveredHFilesOutputSink.java
Undo fetching Table Descriptor. Not reliably available at recovery time.
Signed-off-by: Duo Zhang <zhangduo@apache.org>
After a bit of research into [0] and [1], and a bit of
experimentation, it seems we can use a partial wild-card expression
for these version strings. Let's try this for now. If it works out, we
should expand this usage to all the version package numbers, pinning
them to their epic:upstream-version components.
[0]: http://manpages.ubuntu.com/manpages/xenial/en/man5/deb-version.5.html
[1]: http://manpages.ubuntu.com/manpages/xenial/man8/apt-get.8.html
Signed-off-by: Sean Busbey <busbey@apache.org>
Signed-off-by: Duo Zhang <zhangduo@apache.org>
* put together a docker container that can use host gpg-agent forwarded over ssh.
* use gpg-agent forwarding container on OS X and directly forward the agent on Linux
* clean up the release container on exit
* use docker mounts instead of the deprecated volume syntax
* use image names within our project namespace
* update README to walk through running on GCE with gpg-agent forwarding
closes#1620
Signed-off-by: Matt Foley <mattf@apache.org>
Signed-off-by: Nick Dimiduk <ndimiduk@apache.org>
