[
https://issues.apache.org/jira/browse/HIVE-18263?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16292354#comment-16292354
]
Adam Szita commented on HIVE-18263:
-----------------------------------
Thanks for reviewing [~zsombor.klara], and committing [~pvary]!
This morning after the change was committed I've restarted ptest2, and I've
been monitoring its log:
It started by killing off the existing 12 slaves it was using before, and
simultaneously creating 12 new ones, all according to plan:
{code}
2017-12-15 09:44:41 INFO [localhost-startStop-1] ExecutionController:85 -
Reading configuration from file:
/opt/apache-tomcat-7.0.72/conf/cloudhost.properties
2017-12-15 09:44:46 INFO [localhost-startStop-1]
CloudExecutionContextProvider:130 - CloudExecutionContextProvider
maxHostsPerCreateRequest = 2
2017-12-15 09:44:46 INFO [localhost-startStop-1]
CloudExecutionContextProvider:421 - Requesting termination of [35.225.33.208,
104.198.248.189, 35.184.205.41, 35.184.14.247, 35.192.52.184, 104.154.229.171,
35.192.216.79, 35.224.189.104, 35.184.147.31, 35.225.218.206, 104.198.217.87,
35.224.37.167]
{code}
One hour later the caretaker thread saw the following IP's in
{{mTerminatedHosts}}:
{code}
2017-12-15 10:44:51 INFO [CloudExecutionContextProvider-BackgroundWorker]
CloudExecutionContextProvider:340 - Performing background work
2017-12-15 10:44:51 INFO [CloudExecutionContextProvider-BackgroundWorker]
CloudExecutionContextProvider:345 - Currently tracked terminated hosts:
[35.225.33.208, 35.184.205.41, 104.154.229.171, 35.192.216.79, 35.184.147.31,
35.225.218.206, 35.224.37.167]
{code}
..and it did nothing after, killed no slaves.
All of these 7 IPs can be found in the list of 12 old slaves above (they were
killed at startup). So without this change it is very likely that 5 slaves
would've been killed during test execution unnecessarily.
(e.g. host with IP 104.198.217.87 is not part of mTerminatedHosts, but it is
running right now, and it reuses an IP from the old group)
All-in-all I believe this is working now as it should.
> Ptest execution are multiple times slower sometimes due to dying executor
> slaves
> --------------------------------------------------------------------------------
>
> Key: HIVE-18263
> URL: https://issues.apache.org/jira/browse/HIVE-18263
> Project: Hive
> Issue Type: Bug
> Components: Testing Infrastructure
> Reporter: Adam Szita
> Assignee: Adam Szita
> Fix For: 3.0.0
>
> Attachments: HIVE-18263.0.patch, HIVE-18263.1.patch
>
>
> PreCommit-HIVE-Build job has been seen running very long from time to time.
> Usually it should take about 1.5 hours, but in some cases it took over 4-5
> hours.
> Looking in the logs of one such execution I've seen that some commands that
> were sent to test executing slaves returned 255. Here this typically means
> that there is unknown return code for the remote call since hiveptest-server
> can't reach these slaves anymore.
> In the hiveptest-server logs it is seen that some slaves were killed while
> running the job normally, and here is why:
> * Hive's ptest-server checks periodically in every 60 minutes the status of
> slaves. It also keeps track of slaves that were terminated.
> ** If upon such check it is found that a slave that was already killed
> ([mTerminatedHosts
> map|https://github.com/apache/hive/blob/master/testutils/ptest2/src/main/java/org/apache/hive/ptest/execution/context/CloudExecutionContextProvider.java#L93]
> contains its IP) is still running, it will try and terminate it again.
> * The server also maintains a file on its local FS that contains the IP of
> hosts that were used before. (This probably for resilience reasons)
> ** This file is read when tomcat server starts and if any of the IPs in the
> file are seen as running slaves, ptest will terminate these first so it can
> begin with a fresh start
> ** The IPs of these terminated instances already make their way into
> {{mTerminatedHosts}} upon initialization...
> * The cloud provider may reuse some older IPs, so it is not too rare that the
> same IP that belonged to a terminated host is assigned to a new one.
> This is problematic: Hive ptest's slave caretaker thread kicks in every 60
> minutes and might see a running host that has the same IP as an old slave had
> which was terminated at startup. It will think that this host should be
> terminated since it already tried 60 minutes ago as its IP is in
> {{mTerminatedHosts}}
> We have to fix this by making sure that if a new slave is created, we check
> the contents of {{mTerminatedHosts}} and remove this IP from it if it is
> there.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
