[
https://issues.apache.org/jira/browse/SOLR-15469?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17364163#comment-17364163
]
Danilo commented on SOLR-15469:
-------------------------------
Hi [~mdrob] !
the correct scenario is in the cluster scope. In a cluster like pacemaker every
resource have his start/stop script.
If I configure two or more different solr that differs only for listen ip and
data store...
{code:java}
# /etc/init.d/solr_110 status
No Solr nodes are running.
# /etc/init.d/solr_111 status
No Solr nodes are running.
{code}
{code:java}
# /etc/init.d/solr_110 start
Waiting up to 180 seconds to see Solr running on port 8983 [|]
Started Solr server on port 8983 (pid=4096). Happy searching!
# /etc/init.d/solr_111 status
Found 1 Solr nodes:
Solr process 4096 running on port 8983
Solr at http://192.160.10.111:8983/solr not online.
# echo $?
0
{code}
solr_111 is down but status return as if it were up.
{code:java}
# /etc/init.d/solr_111 start
Port 8983 is already being used by another process (pid: 4096)
Please choose a different port using the -p option.
# ss -tunap | grep 8983
tcp LISTEN 0 50 192.160.10.110:8983 0.0.0.0:* users:(("java",pid=4096,fd=153))
{code}
Then if you run solr_111 on 192.160.10.111:8983 they doesnt' start, wrote the
port is already in use but this is not true.
> start/stop script fails on multiple instance
> --------------------------------------------
>
> Key: SOLR-15469
> URL: https://issues.apache.org/jira/browse/SOLR-15469
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: scripts and tools, SolrCLI
> Reporter: Danilo
> Priority: Minor
> Labels: cluster, instances
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> When more instances of solr run on the same machine. The bash script solr
> fail to get correct pid of process. This cause problems to get status
> information or when stop/restart solr instance.
> In optical to use solr like a cluster service this is essential for the
> cluster to know the healt of service.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
