[
https://issues.apache.org/jira/browse/PHOENIX-5094?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16749577#comment-16749577
]
Kiran Kumar Maturi edited comment on PHOENIX-5094 at 1/23/19 10:11 AM:
-----------------------------------------------------------------------
[~vincentpoon] as suggested I have made the change to increment the counter
when the index transition happens to inactive from disabled.
{code:java}
if (indexState == PIndexState.DISABLE) {
IndexUtil.incrementCounterForIndex(conn, indexTableFullName, 1);
IndexUtil.updateIndexState(conn, indexTableFullName, PIndexState.INACTIVE,
null);
continue; // Must wait until clients start to do index maintenance again
}{code}
and decrement counter for indices that are going to be made active
{code:java}
public List<PTable> decrementIndexesPendingDisableCount(PhoenixConnection conn,
PTable dataPTable, List<PTable> indexes, int amount){
List<PTable> indexesIncremented = new ArrayList<>();
for(PTable index :indexes) {
try {
String indexName = index.getName().getString();
IndexUtil.incrementCounterForIndex(conn, indexName, amount);
indexesIncremented.add(index);
}catch(Exception e) {
LOG.warn("Increment " + amount +" for index :" + index.getName().getString() +
"of table: " + dataPTable.getName().getString() + e);
}
}
return indexesIncremented;
}{code}
i am not checking for the pending disable count == 1 as the count might have
been incremented by temporary failures, which might have been fixed by the
rebuilder.
I have moved PhoeinixIndexFailurePolicy#incrementCounterForIndex to
IndexUtil#incrementCounterForIndex
please review and suggest any other changes to be made. Thanks
was (Author: kiran.maturi):
[~vincentpoon] as suggested I have made the change to increment the counter
when the index transition happens to inactive from disabled.
{code:java}
if (indexState == PIndexState.DISABLE) {
IndexUtil.incrementCounterForIndex(conn, indexTableFullName, 1);
IndexUtil.updateIndexState(conn, indexTableFullName, PIndexState.INACTIVE,
null);
continue; // Must wait until clients start to do index maintenance again
}{code}
and decrement counter for indices that are going to be made active
{code:java}
public List<PTable> decrementIndexesPendingDisableCount(PhoenixConnection conn,
PTable dataPTable, List<PTable> indexes, int amount){
List<PTable> indexesIncremented = new ArrayList<>();
for(PTable index :indexes) {
try {
String indexName = index.getName().getString();
IndexUtil.incrementCounterForIndex(conn, indexName, amount);
indexesIncremented.add(index);
}catch(Exception e) {
LOG.warn("Increment " + amount +" for index :" + index.getName().getString() +
"of table: " + dataPTable.getName().getString() + e);
}
}
return indexesIncremented;
}{code}
i am not checking for the pending disable count == 1 as the count might have
been incremented by temporary failures, which might have been fixed by the
rebuilder.
Good to have changes
1) make PhoenixIndexFailurePolicy#incrementCounterForIndex(old) public or use
IndexUtil#incrementCounterForIndex(new) every where as the function is
currently repeated at. please suggest.
2) Move IndexRebuildIncrementDisableCountIT#mutateRandomly to IndexTestUtil
currently i am using the same code again and again. I will try to find if any
other static methods exist for this please let me know if any better way to do
it.
please review and suggest any other changes to be made. Thanks
> Index can transition from INACTIVE to ACTIVE via Phoenix Client
> ---------------------------------------------------------------
>
> Key: PHOENIX-5094
> URL: https://issues.apache.org/jira/browse/PHOENIX-5094
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 5.0.0, 4.14.1
> Reporter: Monani Mihir
> Assignee: Kiran Kumar Maturi
> Priority: Major
> Attachments: PHOENIX-5094-4.14-HBase-1.3.01.patch
>
>
> Suppose Index is in INACTIVE state and Client load is running continuously.
> With INACTIVE State, client will keep maintaining index.
> Before Rebuilder could run and bring index back in sync with data table, If
> some mutation for Index fails from client side, then client will transition
> Index state (From INACTIVE--> PENDING_DISABLE).
> If client succeeds in writing mutation in subsequent retries, it will
> transition Index state again ( From PENDING_DISABLE --> ACTIVE) .
> Above scenario will leave some part of Index out of sync with data table.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
