Hi
I hope everyone is well and safe.
I was wondering, I have Pure set up to also replicate to BackblazeB2. It works pretty well, but there are often timeouts when uploading the chunks, and so the Upload portion of the job fails. Usually in the log there is an error shown in the event history: Connection Refused or Could Not Queue HTTP Task. I think this is just because B2 is not very reliable. Anyway-
So I was wondering, is there any way to set a retry value on the upload so it doesn't give up as easily? Or, is there any way to schedule an "upload only" job that does not make a new snapshot, but merely tries to make sure the cloud repo has all of the necessary chunks?
Thank you for any answers
Pure already has a mechanism for automatically checking backups and uploading snapshots that are missing. During upload, each chunk is retried several times but if the problem persists, Pure will give up for the time being. In the background, there is an automated process running every 30 minutes which checks all backup snapshots and queues for upload those that still have chunks that need uploading. You can identify this by having Upload jobs in the Event log even at times when backups are not running. Also, those jobs will have fewer chunks to upload. E.g., if the initial job had 2500 chunks but it stopped at 1500/2500, the repeated upload job for the same snapshot will show uploads starting with 1/1000 chunks.
At the moment, there is no way to manually start the upload.
Oh ok thanks, I didn't know that there was a background retry happening. Is this affected by the backup window time at all? Or does it always run 24/7?
It runs all the time. Wakes up every 30 minutes, schedules any pending uploads, goes back to sleep when nothing is left to do.
Got it, thank you for explaining Marijan !
