[
https://issues.apache.org/jira/browse/HDFS-15707?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17245497#comment-17245497
]
Jim Brennan commented on HDFS-15707:
------------------------------------
Thanks for putting this up [~ahussein]. I believe you and [~daryn] worked on
fixing this internally, and I see this includes a related change that was
described by [~daryn]:
{quote}
NNtop is configured to return the top 10 users per-op. This also extends to
the overall total op count. A user may be in the top 10 for individual ops but
not qualify for the overall top 10. This causes nntop to incorrectly graph the
user's ops w/o a total. Instead their top op is confusingly graphed as the
total.
The total op count needs to include all users that appear in the top 10 of any
op.
{quote}
> NNTop counts don't add up as expected
> -------------------------------------
>
> Key: HDFS-15707
> URL: https://issues.apache.org/jira/browse/HDFS-15707
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs, metrics, namenode
> Reporter: Ahmed Hussein
> Assignee: Ahmed Hussein
> Priority: Major
> Labels: pull-request-available
> Time Spent: 40m
> Remaining Estimate: 0h
>
> When we queried
> https://host-name:port/jmx?qry=Hadoop:service=NameNode,name=FSNamesystemState
> earlier today, we received surprising results. Below is a subset of the data.
> The notable thing here is that the count for {{opType='\*'}} from user
> {{user1}} is {{66652}}, but the sum of counts for other {{optype}} values by
> {{user1}} is actually larger: {{77253}}. I'd expect that {{opType='\*'}} to
> always be greater than or equal to the sum for individual operation types.
> {code}
> {
> "windows": [
> {
> "windowLenMs": 300000,
> "ops": [
> {
> "totalCount": 24158,
> "opType": "rpc.complete",
> "topUsers": [{ "count": 2944, "user": "user1" }]
> },
> {
> "totalCount": 15921,
> "opType": "rpc.rename",
> "topUsers": [{ "count": 2891, "user": "user1" }]
> },
> {
> "totalCount": 3015834,
> "opType": "*",
> "topUsers": [{ "count": 66652, "user": "user1" }]
> },
> {
> "totalCount": 2086,
> "opType": "rpc.abandonBlock",
> "topUsers": [{ "count": 603, "user": "user1" }]
> },
> {
> "totalCount": 30258,
> "opType": "rpc.addBlock",
> "topUsers": [{ "count": 3182, "user": "user1" }]
> },
> {
> "totalCount": 101440,
> "opType": "rpc.getServerDefaults",
> "topUsers": [{ "count": 3521, "user": "user1" }]
> },
> {
> "totalCount": 25258,
> "opType": "rpc.create",
> "topUsers": [{ "count": 1864, "user": "user1" }]
> },
> {
> "totalCount": 1377563,
> "opType": "rpc.getFileInfo",
> "topUsers": [{ "count": 56541, "user": "user1" }]
> },
> {
> "totalCount": 60836,
> "opType": "rpc.renewLease",
> "topUsers": [{ "count": 3783, "user": "user1" }]
> },
> {
> "totalCount": 182212,
> "opType": "rpc.getListing",
> "topUsers": [{ "count": 1848, "user": "user1" }]
> },
> {
> "totalCount": 380,
> "opType": "rpc.updateBlockForPipeline",
> "topUsers": [{ "count": 58, "user": "user1" }]
> },
> {
> "totalCount": 215,
> "opType": "rpc.updatePipeline",
> "topUsers": [{ "count": 18, "user": "user1" }]
> }
> ]
> }
> ],
> "timestamp": "2017-07-11T17:20:03+0000"
> }
> {code}
> The full {{TopUserOpCounts}} results from the NNTop endpoint results are
> attached to this JIRA as {{nntop.json}}. The full endpoint output is attached
> as {{endpoint.json}}.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
