SABnzbd Forums

Transitioning from a windows 10 box to a linux media server. Have everything setup on DietPi on an Odroid C2. My windows box used to download at over 20MB/s. I expected a little bit of a hit, but sabnzbd is only pulling 2.5-3MB/s.

I have Multipar on and the latest version of Sabnzbd. What's going wrong and/or how do I start to troubleshoot?

Could be CPU and 'disk' in your 46 USD SBC.

SABnzbd has a CPU and disk speed tester built in: in the SABnzbd GUI, click on the Wrench symbol. Then click on the refresh symbol ... what do you get at "System performance (Pystone)" and "Download folder speed"

My Nanopi Neo2 (an ARM SBC) says 10709 resp 10.5 MB/s.

On a Intel Core with SSD both numbers are 10 times higher...

Could also be sabyenc not being installed ... check your log

I actually think I may have some kind of hardware configuration or kernel problem (I'm a noob with linux so be patient here....) Anyway, I was happy to have Sabnzbd buzzing along at a measly 2.8MB/s and then I fired up QBittorrent and now the entire system is unresponsive. It's not frozen but nothing can load and SSH isn't able to complete before it times out. I need to powercycle it when I'm home and I'll let you know what the results of your test are.

OK, not bad for an ARM box.

Next: check your sanbzbd.log for 'SABYenc'. This is good:

SABYenc module (v3.3.5)... found!

Had to restart it with +Info logging enabled but it does seem to have found it:
INFO::[SABnzbd:404] SABYenc module (v3.3.5)... found!

Any other ideas?

OK, good too.

What speed do you get when you let SAB download this Ubuntu NZB https://www.binsearch.info/?q=ubuntu-16 ... 00&server=

On my Nanopi NEO2 I get an overall speed of 5.3 to 7.2 MB/s: starts at 14 MB/s, drops to 2.x MB/s, and goes up to 3.4 MB/s. I believe this has to do with SAB backthrotling to avoid overloading

I'll test that when I'm able. It's the damndest thing, though -- this box keeps slowing down to almost nothing (SSH freezes up) and then it bounces back for a little while and then it slows down to nothing again. But.... even when it's running "quick" it's still painfully slow running SAB.

Ughhhhh. Froze up for 3 hours and then rebooted. Downloaded your image at 3.4MB/s once it did that. Not sure what my next step is....

How is this device connected to the internet? Hard drive type and connection for both complete and incomplete? Number of connections to the usenet provider?

And check what's using cpu/mem by using 'top'.

Connected with a ethernet cable to my router. Hard drive is an external USB (powered.) Temp drive is running direct to a Sandisk Ultra 128gb. 8 connections to my main provider (with multiple fill providers.)

Here's htop while downloading (all seems well.... just slow?) But notice the server loads in the proceeding time frame. This was when the server became unresponsive. I am not able to capture an htop to see what is going on when that spike happens.


I queued up some files last night and went to bed. Server became unresponsive almost immediately and then stayed that way all night. This morning my services had all restarted after about 10 hours of grinding. Here's the message from SABnzbd:

[FAILED] DietPi-Services | sabnzbd failed (Result: signal) since Wed 2018-05-09 09:22:20 CDT; 1h 1min ago
● sabnzbd.service - sabnzbd DietPi
Loaded: loaded (/etc/systemd/system/sabnzbd.service; disabled; vendor preset: enabled)
Active: failed (Result: signal) since Wed 2018-05-09 09:22:20 CDT; 1h 1min ago
Main PID: 19295 (code=killed, signal=KILL)
CPU: 1h 57min 41.163s

May 09 05:48:49 DietPi python[19295]: 2018-05-09 05:48:49,728::INFO::[assembler:96] Decoding /mnt/dietpi_userdata/downloads/incomplete/---part071.rar yenc
May 09 05:49:20 DietPi python[19295]: 2018-05-09 05:49:20,891::INFO::[directunpacker:265] DirectUnpacked volume 71 for db6d7f7973a14a55ac221669c9ecf035-NZBG
May 09 05:51:51 DietPi python[19295]: 2018-05-09 05:51:51,721::INFO::[assembler:96] Decoding /mnt/dietpi_userdata/downloads/incomplete---/db6d7f7973a14a55ac221669c9ecf035-NZBG.part072.rar yenc
May 09 05:52:26 DietPi python[19295]: 2018-05-09 05:52:26,301::INFO::[directunpacker:265] DirectUnpacked volume 72 for db6d7f7973a14a55ac221669c9ecf035-NZBG
May 09 06:46:49 DietPi python[19295]: 2018-05-09 06:44:40,477::INFO::[downloader:776] Thread [email protected]: timed out
May 09 07:22:15 DietPi python[19295]: 2018-05-09 07:05:06,653::INFO::[downloader:776] Thread [email protected]: timed out
May 09 08:42:40 DietPi python[19295]: 2018-05-09 07:36:20,196::INFO::[downloader:776] Thread [email protected]: timed out
May 09 09:22:20 DietPi systemd[1]: sabnzbd.service: Main process exited, code=killed, status=9/KILL
May 09 09:22:20 DietPi systemd[1]: sabnzbd.service: Unit entered failed state.
May 09 09:22:20 DietPi systemd[1]: sabnzbd.service: Failed with result 'signal'.

Your Odroid C2 has 2GB of RAM. That's nice. Your thop shows 1.68GB available ... which I find a bit strange; where's the rest?

But now: Swap is 331MB, and almost used completely. That is a bad sign; as soon as swap gets used, your system will get slow.

SAB is only using 34% of MEM, so how is the rest used?

Is it possible that dietpi sets up swap on a ramdisk? Seems like an awfully big coincidence if 331mb swap and 1.68gb of ram add up to 2gb total?

Maybe check with "swapon -s"