[
https://issues.apache.org/jira/browse/SOLR-13055?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16716357#comment-16716357
]
Amrit Sarkar commented on SOLR-13055:
-------------------------------------
This can be achieved in multiple manners and following I suggest introducing
"*isLive*" parameter in *SystemInfoHandler*.
Command: {code}http://localhost:8983/solr/admin/info/system?isLive=true{code}
Responses:
{code}
{
"responseHeader":{
"status":0,
"QTime":1524}}
{code}
{code}
{
"responseHeader":{
"status":503,
"QTime":2},
"error":{
"metadata":[
"error-class","org.apache.solr.common.SolrException",
"root-error-class","org.apache.solr.common.SolrException"],
"msg":"Connection to Zookeeper is lost. Node unhealthy.",
"code":503}}
{code}
note: the response metadata is added in every response in the base class.
Three checks for liveliness in place:
1. Node upTime greater than 0 milliseconds.
2. CoreContainer is not NULL and active, not shut.
3. Able to connect to Zookeeper, in case of node started in Solrcloud mode.
I am quite happy with these basic checks at the node level and design and
requesting feedback and suggestions.
> Introduce check to determine "liveliness" of a Solr node
> --------------------------------------------------------
>
> Key: SOLR-13055
> URL: https://issues.apache.org/jira/browse/SOLR-13055
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Amrit Sarkar
> Priority: Minor
> Labels: cloud, native
>
> As the applications are becoming cloud-friendly; there are multiple probes
> which are required to verify availability of a node.
> Like in Kubernetes we need 'liveliness' and 'readiness' probe explained in
> https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-probes/n
> determine if a node is live and ready to serve live traffic.
> Solr should also support such probes out of the box as an API or otherwise to
> make things easier. In this JIRA, we are tracking the necessary checks we
> need to determine if a node is 'liveliness', in all modes, standalone and
> cloud.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
