[
https://issues.apache.org/jira/browse/PHOENIX-1081?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14059499#comment-14059499
]
yang ming edited comment on PHOENIX-1081 at 7/11/14 11:40 PM:
--------------------------------------------------------------
Hbase version is 0.94.19.I had added some debug log,may be the endless loop is
here.
{color:red}at
org.apache.hadoop.hbase.regionserver.StoreScanner.next(StoreScanner.java:390){color}
And the problem may be in this method.
{color:blue}at org.apache.phoenix.filter.SkipScanFilter.navigate{color}
LOOP: while((kv = this.heap.peek()) != null) {
// Check that the heap gives us KVs in an increasing order.
assert prevKV == null || comparator == null ||
comparator.compare(prevKV, kv) <= 0 :
"Key " + prevKV + " followed by a " + "smaller key " + kv + " in cf "
+ store;
prevKV = kv;
{color:red}ScanQueryMatcher.MatchCode qcode = matcher.match(kv);{color}
switch(qcode) {
case INCLUDE:
case INCLUDE_AND_SEEK_NEXT_ROW:
case INCLUDE_AND_SEEK_NEXT_COL:
Filter f = matcher.getFilter();
outResult.add(f == null ? kv : f.transform(kv));
count++;
was (Author: yangming860101):
Hbase version is 0.94.19.I had added some debug log,may be the endless loop is
here.
{color:red}at
org.apache.hadoop.hbase.regionserver.StoreScanner.next(StoreScanner.java:390){color}
And the problem may be in this method.
{color:blue}at org.apache.phoenix.filter.SkipScanFilter.navigate{color}
LOOP: while((kv = this.heap.peek()) != null) {
// Check that the heap gives us KVs in an increasing order.
assert prevKV == null || comparator == null ||
comparator.compare(prevKV, kv) <= 0 :
"Key " + prevKV + " followed by a " + "smaller key " + kv + " in cf "
+ store;
prevKV = kv;
{color:red}
ScanQueryMatcher.MatchCode qcode = matcher.match(kv);
{color}
switch(qcode) {
case INCLUDE:
case INCLUDE_AND_SEEK_NEXT_ROW:
case INCLUDE_AND_SEEK_NEXT_COL:
Filter f = matcher.getFilter();
outResult.add(f == null ? kv : f.transform(kv));
count++;
> CPU usage 100% With phoenix
> ----------------------------
>
> Key: PHOENIX-1081
> URL: https://issues.apache.org/jira/browse/PHOENIX-1081
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 3.0.0
> Reporter: yang ming
> Priority: Critical
> Attachments: JMX.jpg, jstat.jpg, the jstack of all threads, the
> jstack of thread 12725.jpg, the jstack of thread 12748.jpg, the threads of
> regionserver process.jpg
>
>
> The concurrent of the system is not high,but CPU usage often up to 100%.
> I had stopped the system,but regionserver's CPU usage is still high.
> what can case this problem?
> table row count：6000 million
> table ddl：
> create table if not exists summary
> (
> videoid integer not null,
> date date not null,
> platform varchar not null,
> device varchar not null,
> systemgroup varchar not null,
> system varchar not null,
> vv bigint,
> ts bigint,
> up bigint,
> down bigint,
> comment bigint,
> favori bigint,
> favord bigint,
> quote bigint,
> reply bigint
> constraint pk primary key (videoid, date,platform, device, systemgroup,system)
> )salt_buckets = 30,versions=1,compression='snappy';
> query 1：
> select sum(vv) as sumvv,sum(comment) as sumcomment,sum(up) as sumup,sum(down)
> as sumdown,sum(reply) as sumreply,count(*) as count from summary(reply
> bigint) where videoid
> in(137102991,151113895,171559204,171559439,171573932,171573932,171573932,171574082,171574082,171574164,171677219,171794335,171902734,172364368,172475141,172700554,172700554,172700554,172716705,172784258,172835778,173112067,173165316,173165316,173379601,173448315,173503961,173692664,173911358,174077089,174099017,174349633,174349877,174651474,174651474,174759297,174883566,174883566,174987670,174987670,175131298)
> and date>=to_date('2013-09-01','yyyy-MM-dd') and
> date<=to_date('2014-07-07','yyyy-MM-dd')
--
This message was sent by Atlassian JIRA
(v6.2#6252)
