Why is external index rebuilt on restart of docker image in azure app service?
Hi,
we have two sites running in docker on a azure app service, one front server and one backoffice server. Umbraco 13.5.2. Each site has its own fileshare on blob storage to store it's temp files.
I noticed from the logs that when I do a restart of my front server, the externalindex is rebuilt (among others). Can see the result of the healthcheck in logs before it starts rebuilding, "No problems were detected with this index."
Why is the index rebuilt?
Cheers,
Markus
(BTW - these sites could share indexes but since their performance is slow, separating the indexes was something we tested to ensure that there isn't any index conflicts on startup. )
More info from the logs:
"Checked main index {IndexName} and it is clean."
"Replicating index from {SourceIndex} to {DestinationIndex}"
"{IndexName} replication complete from {SourceDirectory} to {DestinationDirectory}"
"Replication from index {SourceIndex} to {DestinationIndex} complete."
"Clearing existing index {IndexName}"
"{IndexName} replication complete from {SourceDirectory} to {DestinationDirectory}"
After looking into more settings, I think it is LuceneDirectoryFactory = SyncedTempFileSystemDirectoryFactory that is causing this behaviour.
The restart will result in a fresh local file system. So indexing files are created in /temp and then replicated because of this setting.
Changing this to Default instead for the front server.
Why is external index rebuilt on restart of docker image in azure app service?
Hi,
we have two sites running in docker on a azure app service, one front server and one backoffice server. Umbraco 13.5.2. Each site has its own fileshare on blob storage to store it's temp files.
I noticed from the logs that when I do a restart of my front server, the externalindex is rebuilt (among others). Can see the result of the healthcheck in logs before it starts rebuilding, "No problems were detected with this index."
Why is the index rebuilt?
Cheers, Markus
(BTW - these sites could share indexes but since their performance is slow, separating the indexes was something we tested to ensure that there isn't any index conflicts on startup. )
More info from the logs: "Checked main index {IndexName} and it is clean." "Replicating index from {SourceIndex} to {DestinationIndex}" "{IndexName} replication complete from {SourceDirectory} to {DestinationDirectory}" "Replication from index {SourceIndex} to {DestinationIndex} complete." "Clearing existing index {IndexName}" "{IndexName} replication complete from {SourceDirectory} to {DestinationDirectory}"
After looking into more settings, I think it is LuceneDirectoryFactory = SyncedTempFileSystemDirectoryFactory that is causing this behaviour. The restart will result in a fresh local file system. So indexing files are created in /temp and then replicated because of this setting.
Changing this to Default instead for the front server.
is working on a reply...