Hi
I have a issue where the index on a loadbalancing setup fails at random times..
Getting this error on the backoffice server and the front-end server
Deleting the temp folder from the windows temp on both servers fixes the issue.
But have seen it happen a couple of times now.
The sites a set to use this:
“LuceneDirectoryFactory”: “TempFileSystemDirectoryFactory”
Anyone know anything about this??
Umbraco is pretty well known for having issues with Examine Indexes in Azure and other container/loadbalanced setups.
But I haven’t really heard of many trouble with direct hosting on Windows with loadbalancing.
However, I think that until the move to the new Umbraco search implementation happens, it’s going to continue being a problem.
As far as I know, then only reliable option is to use ExamineX or Elastic.
But I don’t know anything about implementing or using Elastic.
Load Balancing Azure Web Apps | Umbraco CMS
Just checking you have the recommended approach, eg different settings for your schedulingPublisher vs your subscribers?
And are explicitly setting your schedulingPublisher? (not relying on automatic election)
Hi
Yes, the frontend are subscriber and the backoffice is schedulingPublisher :)
hmm okay, thanks for the response
Actually, if you’re not on the latest version of Examine, then consider upgrading it.
We’ve seen some pretty big improvements by manually upgrading to the latest version.
Could be worth a shot 
Thanks, I will give it a try.
Is there anything on those servers that could be interfering with the file?
AV maybe? Or some well-intentioned clean-up script that automatically removes TEMP files?
Will give it a look, thanks!
Do you know if it’s possible that it don’t saves the index in the windows temp folder?
Ended up using the “Default” looks like that fixed our issue 
