[
https://issues.apache.org/jira/browse/SOLR-2300?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Alexandre Rafalovitch closed SOLR-2300.
---------------------------------------
Resolution: Won't Fix
This is no longer relevant to the current replication method.
> snapinstaller on slave is failing
> ---------------------------------
>
> Key: SOLR-2300
> URL: https://issues.apache.org/jira/browse/SOLR-2300
> Project: Solr
> Issue Type: Bug
> Components: replication (scripts)
> Affects Versions: 1.3
> Environment: Linux, Jboss 5.0GA, solr 1.3.0
> Reporter: sakunthala padmanabhuni
>
> Hi,
> We are using Solr on Mac OSX and it is working fine. Same setup we have
> moved to Linux. We have master, slave setup. Every 5 minutes, index will be
> replicated from Master to Slave and will be installed on slave. But on Linux
> on the slave when the snapinstaller script is called, it is failing and
> showing below error in logs.
> /bin/rm: cannot remove
> `/ngs/app/esearcht/Slave2index/data/index/.nfs000000000001110300000749':
> Device or resource busy
> This error is occuring in snapinstaller script at below lines.
> cp -lr ${name}/ ${data_dir}/index.tmp$$ && \
> /bin/rm -rf ${data_dir}/index && \
> mv -f ${data_dir}/index.tmp$$ ${data_dir}/index
> It is not able to remove the index folder. So the index.tmp files are keep
> growing in the data directory.
> Our data directory is "/ngs/app/esearcht/Slave2index/data". When checked
> with ls -al in the index directory, there are some .nfs files still there,
> which are not letting index directory to be deleted. And these .nfs files
> are still being used by SOLR in jboss.
> This setup is giving issue only in linux. Is this known bug on linux?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
