Connection to host truenas.local is broken

[Phantomsfbw]

I am getting amazing throughput results using Intel X550-T1 NICs and Unifi 10G gear, however large file transfers lock up at around 50-60GBs, and I have to reboot the NAS to do anything else with the files at that point to include trying to finishing copying files or deleting and starting over. The total file size I want to transfer is right at 200GB.

This things is smoking fast, but have not figured out how to nail down the disconnect problem..."Connection to host truenas.local is broken."

Does not matter if I run on windows or linux, same error, same approximate data size. I have not changed any of the Truenas Scale default settings. I am using SABA shares and only one VDEV that is almost 7GBs in size. The four drives are setup in Truenas Raid 10 mode.

My system runs an intel Core i5 10400 with 32GB of Crucial Ram at 3200Mhz on a Gigabyte H470M DSH3 motherboard. The drives are Toshiba N300 4TB.

I also got the same error with 16GBs of RAM. Non of the nightly updates seem to have had an impact either. I did have one test where I manage to copy almost 80GB of data before the connection to host lost message appeared. Ironically, I can move the entire 200GB of files back and forth over a USB connection without issue, but it sure takes a lot longer to accomplish.

Thank you for any insight and hope this helps with testing.

 

[Phantomsfbw]

For fun, just conducted a follow up test with a single SAMSUNG NVME 500GB SSD. Same issue. Get to roughly 50,000 files and the NAS disconnects. I have also noticed that even after a reboot, you have to delete files in batches less than 5 to6000 or the NAS will disconnect. Even if you do them in small batches, the NAS eventually disconnects.

 

[morganL]

It sounds like a bug.. could you report it and provide any debug info so that it can be replicated. Thanks.