very slow download speeds and connection timeouts

[quixotic120]

Hello,

running SABnzbd on a truenas scale server, fresh install, and having an issue where I simply cannot get download speed to go above a crawl. Max I've gotten is 30-40kb/s

Things I've investigated and tried:

resetting network hardware

attempting connection and download of file with newsgroup setting via alternate app on alternate machine - was able to download an NZB with same server settings on my laptop and get significantly better speed, roughly 60mb/s

attempting alternative connections - both utilizing alternative EU server and a backup blocknews account, same issue. also tried with both SSL on and off on all servers

same issue downloading the test file

image showing the pattern of bandwidth too; there is a spiking pattern that occurs which made me think it could be resource related but increasing CPU and memory allocation did not have any impact.

system performance stats in SABnzbd are reported as the following - performance is 576054, bandwidth is 10.9MB/s to the left of the refresh and 87.2mbps to the right, disk speed is reported as 50-90MB/s depending on when I do the scan, sometimes it dips lower.

I've read the disk performance can be the bottleneck here. I removed the nvme l2arc cache from my zfs pool and temporarily created a secondary cache pool for SAB. this reported a disk speed of over 1500MB/s and the issue persisted with no measurable change. I destroyed that pool and readded as cache to the original pool. Additionally qbittorrent on same machine has no issue getting to 40-50mb/s download writing to same pool fwiw. repeated the test download there with all servers, same issue.

logfile, which didnt reveal much to me: pastebin /Z02jGHFf (sorry, cant post as a link because i'm a new user)

any ideas? where can i look from here? thanks

[sander]

and connection timeouts

Can you post the timeouts?

[quixotic120]

so I actually was able to somewhat resolve although I still don't understand why my fix worked or what was wrong. Looking at the log timeouts occurred on random threads and I often got the timeout warning popping up in the connections part of the status pane. So I tried dropping the server connections to 1, restarting the pod, and that worked. then I upped the connections to 5, 10, 15, 20 and it went up although after 20 there was no benefit. Been running for an hour now with no issues.

Now the issue is that my connection speed sucks. it was initially 3MB/s then 11MB/s when i got up to 20 connections but I have gigabit internet. I thought this was disk speed but I tried the nvme pool thing again and that didnt work. further investigation reveals I might have an issue with my truenas setup thats limiting the network bandwidth to 20mbit but that's out of the scope of this forum.

hopefully this post helps someone down the line though. let me know if posting the logs with the timeouts would be helpful

[sander]

Good you have no issues anymore

"truenas scale server," ... is that virtualization, with docker?

[quixotic120]

pretty much, truenas scale is a NAS oriented OS based on debian and it runs apps like SAB virtualized but with kubernetes instead of docker