SABnzbd Forums

I was running sabnzbd on Windows 7 and it was using full 24Mbps but i migrated it to debian, and not I am not getting more than exactly 1.1 Mbps, its like there is a limit set somewhere.

You have an article cache set?
It should run faster on Linux than on Windows (on comparable hardware).

manpreet wrote:24Mbps but i migrated it to debian, and not I am not getting more than exactly 1.1 Mbps, its like there is a limit set somewhere.

Don't you mean 1.1 MB/s?
I got the same issue as you, but I got both IPTV and VOIP that steals bandwidth, or did it go faster when you ran Windows..?

There is not Article cache set. There is nothing else downloading at all. And I mean basically is hould download at 3 Megabytes/s but only doing at 1.1 Megabytes/s

Set an article cache. Currently a lot of time is wasted on disk access.
Set it to 200M if you can afford the memory use.

Allow me to hijack threat for a simple question. I have 4GB of Ram and article cache set to 256MB, running SAB on full speed of 4MB/s. Where is limit of cache size? I mean, should I set it to 512 if I have often many jobs in queue? I don't have any performance issues so is it worth it?

A cache that is large enough to take one complete file (plus a bit extra for the next one)
will reduce the disk access quite a lot.
Generally 200M is enough.
The choice is yours, SABnzbd will function either way.

So I set it up to use 2048MB Article cache (host has 16GB RAM) but its still working on same 1.1 Mbps, I am ready to pull my hair out cant figure where its choking

I understand your frustration, but I cannot tell you where the issue is.
Whether inside or outside SABnzbd.
To add to your misgivings: I run SABnzbd on an old netbook running Ubuntu 8.04
and I can get between 6 and 7 MBytes/sec.
So far all Linux/Unix variants I ran under VMWare have similar or better performance.
What I cannot vary is my router and ISP.

Which Debian do you run exactly?
Did you install the Python-yenc package too?

manpreet wrote:So I set it up to use 2048MB Article cache (host has 16GB RAM) but its still working on same 1.1 Mbps, I am ready to pull my hair out cant figure where its choking

Throttling?
Poor peering?

Sometimes there are random events that affect speed. Good luck.

It issue is it was working perfectly fine when running on Windows.

I do have python-yenc package and using debian testing

Finally fixed the issue apparently kernel 3.1.0 MArvel Yukon drivers auto negotiate at 10Mbps even with 1000mbps cards so to fix

ethtool -s eth0 autoneg off speed 1000 duplex full

Glad you found a solution.
But I hope you understand that it's impossible that we (developers) come up with such a solution.
Sometimes fellow users can suggest such solutions.