Knowledge - Details
Date Added: 7/30/2017
Crashplan limits itself by default to 1GB of RAM. At the start of a backup, the computer will try to scan for all the changes before it starts to send them to the destination machine. If it reaches the RAM limit while scanning, the service will crash and the destination machine will log the backup destination was disconnected. Apparently, Crashplan considers this a successful backup because it never sends you an alert warning you that your backups are not working, even if no backup actually takes place for months. The only way to detect this event is by manually checking the Crashplan destination machine's logs for that error message, watching for service crash events in the Windows Event Viewer on the client machine, or by noticing in your occasional status emails that your percent backed up is far less than 100% even though it claims to have backed up only hours ago. HOW TO FIX THE PROBLEM On the client machine that is crashing, open the program and double click the Crashplan logo. This should bring you to the secret command window. Here, type this command: java mx 2048, restart Replace 2048 with how many MB of RAM you want Crashplan to be allowed to use. Never specify more memory than your computer actually has. You can specify a large number here and Crashplan will only use as much RAM as it actually needs, so it's not wasteful to specify a large number. Remember that the default number is 1024, so 2048 would be doubling that. https://support.code42.com/CrashPlan/5/Troubleshooting/Adjust_CrashPlan_settings_for_memory_usage_with_large_backups
Disclaimer: Everything on this website is written for my own use. I disclaim any guarantees that the procedures and advice listed here are accurate, safe, or beneficial for anyone else. If you attempt to follow any procedures or advice shared here, you do it at your own risk. Part of IT work is knowing how to recover from problems.