PBS Sync job timeout

kriev98

New Member
Feb 14, 2025
15
1
3
I have a new PBS server for an offsite sync job. I'm using WAN because it's 13TB of data and VPN is too slow. I set up my ACL rules and everything however, when i start the sync job, it runs for about 10-15mins then it just times out. I tried lowering MTU but it's still timing out. Anyone have an idea why it's timing out on WAN but on VPN it's fine just way slower.
 
Hi,
have you considered seeding your remote datastore by using a removable datastore? So perform the initial sync from the onsite PBS to the removable datastore, detach the disk and move it to the offsite PBS, add the removable datastore there and then sync to the target datastore there.

Regarding the sync issue: Can you continuously ping the remote while the sync is ongoing? Do you perform a push or pull sync? Please share the exact error messages you get, including some context which might tell at which point the timeout happens.
 
I can't do the removable since the source PBS is in a datacenter, we only rent the server.

I have run a iperf for 1h without any interruptions at full gigabit between both PBS servers.

The configuration is a pull from the remote pbs. The timeout is random, it can be almost immediate and sometimes after a couple of mins.

The only error i see is the source pbs showing a timeout error about 15mins after the data stops flowing and then the remote pbs gets a timeout error about 20mins after.

I also tried playing with TCP timeout settings but nothing is working.
 
ok, i switched back to VPN to get this started but it turns out it times out on VPN too.
 
What Kernel version are your PBS hosts running? Kernel 6.17.2 unfortunately was problematic, it is recommended to use the latest versions available.