Download issues after a week, stability issues after upgrade

[twaddle]

Hi All,

So I've been tinkering with sabnzbd for about a fortnight now, my efforts have been patchy at best and I thought it was finally time to document my problems here to seek some guidance.

First of all, to explain my setup. I've got an AMD box with 8GB RAM running unRAID, I'm aware that unRAID+SAB can have some issues when the storage array is used, so I opted to follow what another user did here http://www.lockstockmods.net/2010/06/03 ... sickbeard/ and got sab/sickbeard working on my cache drive, which worked pretty well for about a week.

This is the data for that setup:

Version: 0.5.2
OS: unRAID (Slackware Derived)
Install-type: python source (server is running python 2.5.2 (r252:60911, Sep 11 2008, 13:43:31) [GCC 4.2.4])
Skin (if applicable): Plush
Firewall Software: No firewall on the unraid, NAT firewall on edge
Are you using IPV6? no
Is the issue reproducible? yes

The only big issue with this setup was that downloads would occasionally fail yet seemed to extract fine when I used winRAR. Then, about a week ago when every download would routinely stall. I read around, came to IRC, cleared cache folders and did various other things I found on the forums but couldn't find a soultion. Each time SAB started my bandwidth spiked at about 20Mb for a second or two, then tapered off to 0. I'm using Virgin Media 50Mb, Astraweb, SSL on port 443.

So, I thought before I'd come ask for help that I'd update my box to the latest, 0.6.14:

Version: 0.6.14
OS: unRAID (Slackware Derived)
Install-type: python source (server is running python 2.5.2 (r252:60911, Sep 11 2008, 13:43:31) [GCC 4.2.4])
Skin (if applicable): Plush
Firewall Software: No firewall on the unraid, NAT firewall on edge
Are you using IPV6? no
Is the issue reproducible? yes

In this case transfers began working again, although periodically and without reason SAB just seems to stop running, I can't find anything in the logs with debugging enabled which explains it, however if I remove any incomplete downloads it seems to stay up for longer than if I don't.

In the case of downloads which do succeed on 0.6.14, something strange seems to happen with the sabtosickbeard.py, entire unpacked folders get moved instead of just the media file inside and they're binned in the root folder for that type of content rather than the subfolder they're intended for. Appreciate this might be a sickbeard issue but thought I'd mention it here in case I've misinterpreted.

Any ideas folks?

[shypike]

Please read this and check if that solves your problem.
http://forums.sabnzbd.org/viewtopic.php?f=3&t=9708

The essence is that there seems to an issue with one of the Python libraries.

[twaddle]

Awesome, thanks, that might explain the crashing, any ideas what else might have changed for the sickbeard script to be misbehaving?

[shypike]

No ideas about the SB script.
Nothing changed in script support for quite some time.

[twaddle]

I've successfully upgraded to 0.6.14 now, everything works fine with one exception

I've made the tweak re pickle/cPickle and hoped it would cure stability, but it hasn't.

In the Syslog, I can see the following each time sabnzbd falls over:

kernel: python[7651]: segfault at 18 ip b77d6b74 sp a4ffe420 error 4 in libpython2.5.so.1.0[b771f000+10a000]

This has only begun on 0.6.14, if I rever to 0.5.20 it's peachy (although the download issue on 0.5.20 persists).

Any ideas?

[shypike]

Doesn't going back to 0.6.9 solve the issue?

[twaddle]

I've not used it, I'll grab a copy and give it a whirl.

[shypike]

Is this a 64bit OS and possibly a 64bit Python port?

[twaddle]

It's unRAID, the hardware is x64 capable but I honestly don't know whether the distribution is, python details are:

Python Version: 2.5.2 (r252:60911, Sep 11 2008, 13:43:31) [GCC 4.2.4]

Don't see any mention of x64.

[twaddle]

ok, 0.6.9 just fell over too:

kernel: python[14328]: segfault at ae ip b765c6b6 sp a6dfe6b0 error 4 in libpython2.5.so.1.0[b75ef000+10a000]

Did something more fundamental relating to python change between 0.5.x and 0.6.x? :-(

[twaddle]

Upgraded Python, now running:

Python Version: 2.6.5 (r265:79063, Jul 24 2010, 21:39:31) [GCC 4.2.3]

Will re-test.

[twaddle]

Python upgrade has resolved it, does that raise the minimum Python version required in 0.6.x (or did I miss that documented elsewhere already?)

[shypike]

It isn't so much the Python version that matters, we do support Python 2.5.
However there are many 2.5 sub-releases and also not each port is of equal quality.
Anyway, good that it works now.