[
https://issues.apache.org/jira/browse/HBASE-1150?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nitay Joffe updated HBASE-1150:
-------------------------------
Attachment: hbase-1150.patch
The main logical change of this patch is this:
--- a/src/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
+++ b/src/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
@@ -380,7 +380,7 @@ public class HRegionServer implements HConstants,
HRegionInterface, HBaseRPCErro
i++) {
LOG.info(msgs[i].toString());
if (safeMode.get()) {
- if (!msgs[i].isInSafeMode()) {
+ if (zooKeeperWrapper.checkOutOfSafeMode()) {
this.connection.unsetRootRegionLocation();
synchronized (safeMode) {
safeMode.set(false);
Everything else is just removing safeMode member from HMsg.
> HMsg carries safemode flag; remove
> ----------------------------------
>
> Key: HBASE-1150
> URL: https://issues.apache.org/jira/browse/HBASE-1150
> Project: Hadoop HBase
> Issue Type: Sub-task
> Reporter: stack
> Assignee: Nitay Joffe
> Attachments: hbase-1150.patch
>
>
> HMsg got a state flag on committ of HBASE-1121. This kinda pollutes the
> simple HMsg type -- a shell for carrying state as byte type -- to instead
> being a vehicle that carries state in byte and in a boolean that is only used
> on cluster start then never read again but is passed in all HMsgs anyways.
> Can we get rid of it when we integrate zk? Go to zk to figure if master is
> in safe mode?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
