Kiran Kumar Dasari
Kiran Kumar Dasari

Reputation: 11

SOLR replication failure on one of the slaves

We have a 4 slaves on Windows 2008 environment on tomacat and replication was working fine for some time and it started failing on one of the nodes i see following errors at first look it looks like permission issue but i compared two nodes and they seem to be same and nothing changed on this node.

SEVERE: SnapPull failed 
org.apache.solr.common.SolrException: Unable to rename: D:\solr\core0\conf\compoundwords-de.txt to: D:\solr\core0\conf\compoundwords-de.txt.20120703165100
SEVERE: SnapPull failed 
org.apache.solr.common.SolrException: Failed to create temporary config folder: conf.20120705004320

I even tried restarting the node to remove any pending locks but it did not resolve the issue anything i can do to troubleshoot the issue and find the real cause.

Upvotes: 0

Views: 3174

Answers (2)

arun
arun

Reputation: 11023

I was facing the same issue:

SEVERE: SnapPull failed org.apache.solr.common.SolrException: Failed to create temporary config folder: conf.20120705004320

I happened to notice this error in tomcat's catalina.out after trying to figure out why the Solr admin UI is showing files being transferred during replication, but the index version and gen on the slave do not get updated after replication. In fact, in my case, the slave's version and gen were higher than what were on master (Solr 4.2.1)!

The trouble was the owner of the parent dir of solr conf dir (the one containing schema.xml, solrconfig.xml. etc.,). Solr wants to create a temporary conf dir named like conf.20120705004320 exactly at the same place where conf dir is located.

Once I changed the owner of the parent folder to tomcat6 replication started working fine. I used the command chown -R tomcat6:tomcat6 /var/solr where /var/solr is my ${solr.home}. The slave's version and gen started following the masters' after this fix.

Upvotes: 2

Kiran Kumar Dasari
Kiran Kumar Dasari

Reputation: 11

I finally go this resolved after some troubleshooting we found that there was a failed java update since then the replication started failing. Here is what we did to resolve this

  1. Reinstall JRE again
  2. Tried restarting Solr, and removed all index files but it did not work
  3. We went ahead deleted the core and setup the core again and it started working like before

Upvotes: 1

Related Questions