What is actively limiting the speed of the backup? Not asking about the setting really, just the mechanism behind it. Reason being I have one client being backed up that seems to limit itself to 2mbps though the network is more than capable of higher speeds. Oddly, it can backup 3 drives at that speed concurrently. In fact, I’ve seen the same client come across at about 25mbps months ago before I noticed this issue. I have double checked BWM/QOS on the client network and run speedtests and such. Also tried reinstalling the urbackup client.
Is it at all possible for this to be a stuck registry entry or something? This client has a large data set so I do not want to remove it from the server.
Here’s what I’ve done to troubleshoot but have not gotten it resolved:
-turn off background priority, encryption, and compression
-completely remove client and all backup data from server, uninstall client, and reinstall
-change internet port of server (used port 570 temporarily to test)
The client that is having speed issues has an https server- I can download a file through a web browser on the urbackup server from the slow client at about 25mbps. Urbackup caps at about 2.2mbps though, averages 1.78mbps.
I should also mention that I’m using a cbt license on this client (helps with their big data set of 1TB.)
I know I’m typing into the dark here, but hopefully I can get this resolved and it will help someone else in the future. Here are additional steps I have taken:
-uninstalled the cbt client and installed client from urbackup.org (version 2.2.5)
-installed client on another machine on the remote LAN, which had full speed backup of about 25mbps
-reinstalled chipset and storage drivers on client
Still no dice! Image backups run at about 1.8-2mbps.
Maybe onto something here- I’ve disabled the “background priority” checkbox, however it is still running as background in the Windows Performance Monitor IO priority.
At first glance packet captures with Wireshark seemed fine… until I noticed that nearly every packet had a ECN tag on it. Turns out this is due to CWR, congestion window reduced, throttling the traffic for some reason.
This command resolved the issue, after rebooting the BACKUP server (which I thought was odd) and not the CLIENT.
netsh interface tcp set global ecncapability=disabled
