Help with slow Time Machine Speeds

[djseto]

I've been running Free....err...TrueNAS (Core) for close to 10 years now for home use. It's been largley set and forget except for software updates and replacing bad drives. I noticed lately my Time Machine backups from the Macs in my house are running crazy slow. When I look at the Interface Stats, I see like 190-200KB/s speeds my ingress traffic. If I were to mount a file share (AFP protocol) from my Mac and do a straight drag and drop file transfer, I get upwards of 38-40MB/s transfer rate ingress and egress.

Any ideas on where to start?

The hardware is def old but it's an AMD Athlon II X2 250 Processor with 8GB RAM with Intel 1GB network card in the latest GA version of TrueNAS çore. I have 5 HDs (non SSD) in a RAIDZ2 config. I have a 120 GB SSD that I use to boot from. It's been so long that I don't know if I could/did configure the SSD as a cache (can you do that if OS boots from it)?

 

[anodos]

Maybe verify that you are not running out of space. If there's no aesni support for your processor then you may want to turn off signing client-side (on MacOS). Ahh... AFP... (signing advice may not apply) make sure you also don't have a dataset quota that's close to being exceeded on the time machine target.

 

[djseto]

my pool is at 81% utilization. I have one pool with multiple datasets which include a dataset for backups.

Pool1 (81%
|---> FileShare
|-->MacBackups1 (TimeMachine)
|-->MacBackups2 (TimeMachine)
|-->MacBackups...(TimeMachine)

I also noticed this in my shell

Code:

freenas kernel: Limiting closed port RST response from 1697 to 200 packets/sec

 

[djseto]

Solved. I didn't realize Apple was deprecating AFP. I created a new dataset from scratch and shared it using SMB and it is fast now!

 

[anodos]

Solved. I didn't realize Apple was deprecating AFP. I created a new dataset from scratch and shared it using SMB and it is fast now!

Yeah, we've removed AFP support in SCALE. The protocol in general is on its way out.