[
https://issues.apache.org/jira/browse/SOLR-11456?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hoss Man updated SOLR-11456:
----------------------------
Attachment: SOLR-11456.patch
here's a patch with an attempted fix using {{expectThrows}} but on my machine
it fails -- no RemoteSolrException is thrown.
best guess: the reconnect logic seems to be happening very quickly, so maybe
the node thinks it's healthy again by the time the healthcheck call is made?
{noformat}
[junit4] 2> 1780 INFO
(TEST-HealthCheckHandlerTest.testHealthCheckHandler-seed#[D95E86B813E36BF4]) [
] o.a.s.SolrTestCaseJ4 ###Starting testHealthCheckHandler
[junit4] 2> 2020 INFO (qtp1914125569-26) [n:127.0.0.1:52504_solr ]
o.a.s.s.HttpSolrCall [admin] webapp=null path=/admin/health
params={wt=javabin&version=2} status=0 QTime=22
[junit4] 2> 2055 INFO
(TEST-HealthCheckHandlerTest.testHealthCheckHandler-seed#[D95E86B813E36BF4]) [
] o.a.s.c.MiniSolrCloudCluster Expired zookeeper session 98802019973332995
from node http://127.0.0.1:52504/solr
[junit4] 2> 2056 WARN
(jetty-launcher-1-thread-1-SendThread(127.0.0.1:53478)) [n:127.0.0.1:52504_solr
] o.a.z.ClientCnxn Session 0x15f03e9e79a0003 for server
127.0.0.1/127.0.0.1:53478, unexpected error, closing socket connection and
attempting reconnect
[junit4] 2> java.nio.channels.ClosedChannelException
[junit4] 2> at
sun.nio.ch.SocketChannelImpl.ensureReadOpen(SocketChannelImpl.java:257)
[junit4] 2> at
sun.nio.ch.SocketChannelImpl.read(SocketChannelImpl.java:300)
[junit4] 2> at
org.apache.zookeeper.ClientCnxnSocketNIO.doIO(ClientCnxnSocketNIO.java:68)
[junit4] 2> at
org.apache.zookeeper.ClientCnxnSocketNIO.doTransport(ClientCnxnSocketNIO.java:366)
[junit4] 2> at
org.apache.zookeeper.ClientCnxn$SendThread.run(ClientCnxn.java:1141)
[junit4] 2> 2057 INFO (zkCallback-13-thread-1) [ ]
o.a.s.c.c.ZkStateReader Updated live nodes from ZooKeeper... (1) -> (0)
[junit4] 2> 2159 INFO (qtp1914125569-28) [n:127.0.0.1:52504_solr ]
o.a.s.s.HttpSolrCall [admin] webapp=null path=/admin/health
params={wt=javabin&version=2} status=0 QTime=0
[junit4] 2> 2160 WARN
(zkCallback-8-thread-1-processing-n:127.0.0.1:52504_solr)
[n:127.0.0.1:52504_solr ] o.a.s.c.c.ConnectionManager Watcher
org.apache.solr.common.cloud.ConnectionManager@614458f5 name:
ZooKeeperConnection Watcher:127.0.0.1:53478/solr got event WatchedEvent
state:Disconnected type:None path:null path: null type: None
[junit4] 2> 2160 WARN
(zkCallback-8-thread-1-processing-n:127.0.0.1:52504_solr)
[n:127.0.0.1:52504_solr ] o.a.s.c.c.ConnectionManager zkClient has
disconnected
[junit4] 2> 2162 INFO
(TEST-HealthCheckHandlerTest.testHealthCheckHandler-seed#[D95E86B813E36BF4]) [
] o.a.s.SolrTestCaseJ4 ###Ending testHealthCheckHandler
[junit4] 2> NOTE: reproduce with: ant test
-Dtestcase=HealthCheckHandlerTest -Dtests.method=testHealthCheckHandler
-Dtests.seed=D95E86B813E36BF4 -Dtests.multiplier=2 -Dtests.slow=true
-Dtests.locale=es-EC -Dtests.timezone=AGT -Dtests.asserts=true
-Dtests.file.encoding=UTF-8q
[junit4] FAILURE 0.41s | HealthCheckHandlerTest.testHealthCheckHandler <<<
[junit4] > Throwable #1: junit.framework.AssertionFailedError: Expected
exception RemoteSolrException but no exception was thrown
[junit4] > at
__randomizedtesting.SeedInfo.seed([D95E86B813E36BF4:F84CC1A626ECEF60]:0)
[junit4] > at
org.apache.lucene.util.LuceneTestCase.expectThrows(LuceneTestCase.java:2672)
[junit4] > at
org.apache.solr.cloud.HealthCheckHandlerTest.testHealthCheckHandler(HealthCheckHandlerTest.java:69)
[junit4] > at java.lang.Thread.run(Thread.java:748)
{noformat}
> HealthCheckHandlerTest.testHealthCheckHandler() asserts nothing useful
> ----------------------------------------------------------------------
>
> Key: SOLR-11456
> URL: https://issues.apache.org/jira/browse/SOLR-11456
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Hoss Man
> Attachments: SOLR-11456.patch
>
>
> The meat of {{HealthCheckHandlerTest.testHealthCheckHandler()}} currently
> looks like this...
> {code}
> while (live_nodes.size() == 1 && counter++ < 100) {
> Thread.sleep(100);
> live_nodes =
> cluster.getSolrClient().getZkStateReader().getClusterState().getLiveNodes();
> }
> try {
> req.process(httpSolrClient);
> } catch (HttpSolrClient.RemoteSolrException e) {
> assertTrue(e.getMessage(), e.getMessage().contains("Host
> Unavailable"));
> assertEquals(SolrException.ErrorCode.SERVICE_UNAVAILABLE.code,
> e.code());
> }
> {code}
> * nothing actually asserts that {{live_nodes.size() != 1}} -- so the loop
> could just exit after 100 iterations w/o any changes to hte system happening.
> * instead of using expectThrows, the only assertions are inside the catch
> block -- if the code (unexpectedly) succeeds the test will not fail (false
> positive)
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
