Hello,
We are running Solr 8.2.0 in cloud mode with a 3 node cluster in Azure
kubernetes. We have an Azure fileshare mounted in order to perform backups
using the collection API e.g., /mnt/azure
The collection I'm using to test has 1 shard and 3 replicas with about 17GB
worth of data per replica. While testing backups using the following it
seems to be working for a bit (there is a snapshot directory for the shard
created in the fileshare) but then it soon fails (Note before testing each
time I've ensure the directory, TestLargeCollection, no longer exists);
e.g.,
/solr/admin/collections?action=BACKUP&name=TestLargeCollection&location=/mnt/azure&collection=TestLargeCollection&async=1000
/solr/admin/collections?action=REQUESTSTATUS&requestid=1000&wt=xml
<?xml version="1.0" encoding="UTF-8"?>
<response>
<lst name="responseHeader">
<int name="status">0</int>
<int name="QTime">14</int>
</lst>
<str name="Operation backup caused
exception:">org.apache.solr.common.SolrException:org.apache.solr.common.SolrException:
The backup directory already exists:
file:///mnt/azure/TestLargeCollection/</str>
<lst name="exception">
<str name="msg">The backup directory already exists:
file:///mnt/azure/TestLargeCollection/</str>
<int name="rspCode">400</int>
</lst>
<lst name="status">
<str name="state">failed</str>
<str name="msg">found [1000] in failed tasks</str>
</lst>
</response>
The directory contents looks like this after it fails (notice the missing
backup.properties file);
solr@solr-0:/mnt/azure/TestLargeCollection$ ls
snapshot.shard1 zk_backup
FWIW If I do the same with a smaller collection (only about 1GB) it seems to
work fine;
solr@solr-0:/mnt/azure/TestSmallCollection$ ls
backup.properties snapshot.shard1 zk_backup
Any information would be greatly appreciated.
--
Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html
