Skip to content

HBASE-24675: On Master restart all servers are assigned to default rs… #2053

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
arshadmohammad wants to merge 1 commit into from
Closed

HBASE-24675: On Master restart all servers are assigned to default rs… #2053

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions hbase-server/src/main/java/org/apache/hadoop/hbase/rsgroup/RSGroupInfoManagerImpl.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -591,7 +591,7 @@ private synchronized void refresh(boolean forceOnline) throws IOException {

// This is added to the last of the list so it overwrites the 'default' rsgroup loaded
// from region group table or zk
groupList.add(new RSGroupInfo(RSGroupInfo.DEFAULT_GROUP, getDefaultServers()));
groupList.add(new RSGroupInfo(RSGroupInfo.DEFAULT_GROUP, getDefaultServers(groupList)));
Copy link
Contributor

@virajjasani virajjasani Jul 13, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

refresh fetching info from rsgroup table / ZK makes sense.

Copy link
Contributor

@anoopsjohn anoopsjohn Jul 20, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So the issue is by this time, the in memory state 'holder' not created in the HM restart case. This is a very bad bug.. This should go into all active branches no?

Copy link
Contributor Author

@arshadmohammad arshadmohammad Jul 20, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

branch-1 fix was pending, raised PR #2102


// populate the data
HashMap<String, RSGroupInfo> newGroupMap = Maps.newHashMap();
Expand Down Expand Up @@ -726,9 +726,14 @@ private void updateCacheOfRSGroups(final Set<String> currentGroups) {

// Called by ServerEventsListenerThread. Presume it has lock on this manager when it runs.
private SortedSet<Address> getDefaultServers() {
return getDefaultServers(listRSGroups()/* get from rsGroupMap */);
}

// Called by ServerEventsListenerThread. Presume it has lock on this manager when it runs.
Copy link
Contributor

@anoopsjohn anoopsjohn Jul 20, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also pls leave this comment. The ServerEventsListenerThread calls the above method only.

Copy link
Contributor Author

@arshadmohammad arshadmohammad Jul 20, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

right, comment is not relevant on getDefaultServers(List rsGroupInfoList) method.
but this PR is already merged. shall i create new PR

private SortedSet<Address> getDefaultServers(List<RSGroupInfo> rsGroupInfoList) {
// Build a list of servers in other groups than default group, from rsGroupMap
Set<Address> serversInOtherGroup = new HashSet<>();
for (RSGroupInfo group : listRSGroups() /* get from rsGroupMap */) {
for (RSGroupInfo group : rsGroupInfoList) {
if (!RSGroupInfo.DEFAULT_GROUP.equals(group.getName())) { // not default group
serversInOtherGroup.addAll(group.getServers());
}
Expand Down