If an unassign is unable to communicate with its target server,
expire the server and then wait on a signal from ServerCrashProcedure
before proceeding. The unassign has lock on the region so no one else
can proceed till we complete. We prevent any subsequent assign from
running until logs have been split for crashed server.
In AssignProcedure, do not assign if table is DISABLING or DISABLED.
M hbase-server/src/main/java/org/apache/hadoop/hbase/master/assignment/RegionTransitionProcedure.java
Change remoteCallFailed so it returns boolean on whether implementor
wants to stay suspended or not.
M hbase-server/src/main/java/org/apache/hadoop/hbase/master/assignment/UnassignProcedure.java
Doc. Also, if we are unable to talk to remote server, expire it and
then wait on SCP to wake us up after it has processed logs for failed
server.
Added optional (fourth) parameter "server-principal"
The solution is backward compatible, in case not given, uses "hbase" as default value
If the third parameter is skipped the fourth cannot be set.
Signed-off-by: Josh Elser <elserj@apache.org>
This test as it is written currently will not work with AMv2. This needs to be re-written after HBASE-18511 is committed. Disabled the test and update JIRA to re-enable it with dependency on HBASE-18511.
Signed-off-by: Michael Stack <stack@apache.org>
* Modified AssignmentManager.processAssignQueue() method to consider only highest versioned region servers for system table regions when
destination server is not specified for them. Destination server is retained, if specified.
* Modified MoveRegionProcedure to allow null value for destination server i.e. moving a region from specific source server to non-specific/ unknown
destination server (picked by load-balancer) is supported now.
* Removed destination server selection from HMaster.checkIfShouldMoveSystemRegionAsync(), as destination server will be picked by load balancer
Signed-off-by: Michael Stack <stack@apache.org>
Contains the following commits:
HBASE-17748 Include HBase snapshots in space quotas
Introduces a new Chore in the Master which computes the size
of the snapshots included in a cluster. The size of these
snapshots are included in the table's which the snapshot was created
from HDFS usage.
Includes some test stabilization, trying to make the tests more
deterministic by ensuring we observe stable values as we know
that those values are mutable. This should help avoid problems
where size reports are delayed and we see an incomplete value.
HBASE-17752 Shell command to list snapshot sizes WRT quotas
HBASE-17840 Update hbase book to space quotas on snapshots
- Stop presuming job will be configured with checkout into a 'component' dir
- Disable autocheckout
- Specify checkout within 'component' dir
- leave the workspace around on pipeline failure so we can browse it.
* Jenkinsfile that works for all current branches.
* adds dev-support script for setting parameters for our yetus nightly invocation
* quiets the "zip test results" step
Signed-off-by: tedyu <yuzhihong@gmail.com>
- Converted to python
- Updated to most recent release of Java ACC
- Added ability to include known number of problems
Author: Andrew Wang
Amending-Author: Alex Leblang
Amending-Author: Sean Busbey
Signed-off-by: Dima Spivak <dimaspivak@apache.org>
Signed-off-by: Sean Busbey <busbey@apache.org>
Other changes:
- Update corresponding tests in TestAdmin2. Removed tests centered around serverName part of old functions.
- Remove dead functions from ProtobufUtil and ServerManager
- Rename closeRegion* functions in HBTU to unassignRegion*
Change-Id: Ib9bdeb185e10750daf652be0bb328306accb73ab
Selective add of dependency on hbase-thirdparty jars.
Update to READMEs on how protobuf is done (and update to refguide).
Removed all checked in generated protobuf files. They are generated
on the fly now as part of mainline build.
This patch removes start(MasterProcedureEnv) from
ServerCrashProcedure.java which was a misnomer as a no-op. It
did not start anything.
Change-Id: I4e91864ace912e137471bfce03516746c4aff83e
Signed-off-by: Michael Stack <stack@apache.org>
unbalance.kill.meta.rs property was added which controls the monkey to
kill that region server which holds hbase:meta.
Change-Id: I049d05cc13a78ba868c917879f67ef4f34fffbf9
Signed-off-by: Apekshit Sharma <appy@apache.org>
Michael Stack
Guanghao Zhang
Yu Li
Tamas Penzes
Esteban Gutierrez
Umesh Agashe
dongtao.zhang
Ashu Pachauri
zhangduo
Andrew Purtell
Vikas Vishwakarma
Chia-Ping Tsai
Josh Elser
Sean Busbey
no_apologies
Zach York
Alex Leblang
Benedict Jin
Robert Yokota
Apekshit Sharma
Konstantin Ryakhovskiy
James Taylor
Abhishek Singh Chouhan
Yi Liang
anoopsamjohn
Balazs Meszaros