[ 
https://issues.apache.org/jira/browse/SOLR-13159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16756732#comment-16756732
 ] 

Gus Heck commented on SOLR-13159:
---------------------------------

DNS resolution issues have been confirmed in this case. /etc/hosts had been 
manipulated but the machines were not preserving their identity across 
shut-down & startup. The fix was to set the preserve_hostname: true in 
/etc/cloud/cloud.cfg (AWS feature). So /etc/hosts is where the false memory of 
old ip addresses was. Not a fault in solr, but very confusing. Perhaps if there 
seems to be a relevant segment of the ref guide or wiki a caution note could be 
added, but only if there seems to be somewhere appropriate. 

One tiny tidbit left here might be to track down how the DNS issue allowed for 
the misaligned core count vs replica list. Sounds like we have two sources for 
the same info (that then disagreed when DNS was a mess).

> Autoscaling not distributing collection evenly
> ----------------------------------------------
>
>                 Key: SOLR-13159
>                 URL: https://issues.apache.org/jira/browse/SOLR-13159
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: AutoScaling
>    Affects Versions: 8.0
>            Reporter: Gus Heck
>            Priority: Major
>         Attachments: autoscaling.json, clstat.json
>
>
> I recently ran into a very strange behavior described in detail in the mail 
> linked at the bottom of this description. In short: 
>  # Default settings didn't distribute nodes evenly on brand new 50 node 
> cluster
>  # Can't seem to write rules producing suggestions to distribute them evenly 
>  # Suggestions are made that then fail despite quiet cluster, no changes.
> Also of note was diagnostic output containing this seemingly impossible 
> result with 2 cores counted and no replicas listed:
> {code:java}
>             {
>                 "node": "solr-2.customer.redacted.com:8983_solr",
>                 "isLive": true,
>                 "cores": 2,
>                 "freedisk": 140.03918838500977,
>                 "totaldisk": 147.5209503173828,
>                 "replicas": {}
>             },{code}
> I will attach anonymized cluster status output and autoscaling.json shortly 
> This issue may be related to SOLR-13142
> http://mail-archives.apache.org/mod_mbox/lucene-dev/201901.mbox/%3CCAEUNc48HRZA7qo-uKtJQEtZnO9VG9OErQZGzoOmCTBe7C9zvNw%40mail.gmail.com%3E
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to