Download queue gets "corrupted"

[Senaxx]

Hello,

I'm using SABNZBD 0.7.3 on a QNAP TS 112. Normally when I download everything immediately everything works ok. Good download speed and everything works well. But on this QNAP I also use a scheduler. And sometimes I have the problem that it "corrupts" the download queue. When the schedule starts it tries to start downloading the download that I added but, somehow it can't find the articles. When I just restart SABNZBD, it still fails to download, but when I choose for "download and repair" and SABNZBD comes back up it starts the download successfully. But the problem doesn't persist all the time.

So i'm trying to figure out on what condition sabnzbd failes to download the articles. i've looked in the log, but I can't see any errors. Could this be due memory limit on the TS-112? Furthermore there is only Sickbeard running on the QNAP, so not really memory hungry programs.

Any hints?

[shypike]

How much memory do you have?
Some systems have a bug in their Python implementation, requiring an option in Config->Special.
Set option "use_pickle" on (tick the box), click Save and restart SABnzbd.

[Senaxx]

The problem still persist, and now also occured on my own NAS.

QNAP 1: TS -112 -> 256 MB
QNAP 2: TS - 559 -> 1 GB

Main difference is that QNAP 1 works with a scheduler, and QNAP 2 doesn't. Furthermore it runs 0.7.3 as latest availbe SABNZBD, and i can see that it ALWAYS solves the problems when I do a repair and restart, just a restart won't fix the problem. I've tried your setting on both QNAP's and this didn't solve the problem.

Allthough i saw some similair issue here on the forum and that's fixed in 0.7.4. Maybe when i'll upgrade this will get fixed.

[shypike]

256M is quite low, too low for comfort actually.

[sander]

The TS -112 -> 256 MB system: has it got swap space? If so, how much?

You can check swap space it two ways. For both ways you have to login on your NAS device running SABnzbd

Unix in general (I hope):

Code: Select all

pi@raspberrypi ~ $ top -bn1 | head -6
top - 21:18:00 up 3 days, 23:52,  1 user,  load average: 1.17, 1.07, 1.06
Tasks:  62 total,   1 running,  61 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.3 us,  0.1 sy, 99.6 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:    189104 total,   140476 used,    48628 free,    13356 buffers
KiB Swap:   102396 total,     1144 used,   101252 free,    94560 cached

Method 2 (Linux only):

Code: Select all

pi@raspberrypi ~ $ cat /proc/swaps 
Filename				Type		Size	Used	Priority
/var/swap                               file		102396	1144	-1
pi@raspberrypi ~ $

Can you post the output?

[Senaxx]

256M is quite low, too low for comfort actually.

Problem also persist on my 1GB system, and there is a 3rs TS 112 that doesn't have the problem.
1 common thing, they all share the same newsserver. Supernews which is allowed 30 connection from 5 ip's. 2x TS-112 systems use 4, and TS-559 uses 12.

The TS -112 -> 256 MB system: has it got swap space? If so, how much?

You can check swap space it two ways. For both ways you have to login on your NAS device running SABnzbd

Unix in general (I hope):

Code: Select all

pi@raspberrypi ~ $ top -bn1 | head -6
top - 21:18:00 up 3 days, 23:52,  1 user,  load average: 1.17, 1.07, 1.06
Tasks:  62 total,   1 running,  61 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.3 us,  0.1 sy, 99.6 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:    189104 total,   140476 used,    48628 free,    13356 buffers
KiB Swap:   102396 total,     1144 used,   101252 free,    94560 cached

Method 2 (Linux only):

Code: Select all

pi@raspberrypi ~ $ cat /proc/swaps 
Filename				Type		Size	Used	Priority
/var/swap                               file		102396	1144	-1
pi@raspberrypi ~ $

Can you post the output?

Output from the the swap

Code: Select all

[~] # cat /proc/swaps
Filename                                Type            Size    Used    Priority
/dev/md1                                partition       530040  38664   -1

Top command is not working, so i've used normal top:

Code: Select all

Mem: 232580K used, 23048K free, 0K shrd, 3956K buff, 90548K cached
Load average: 0.60, 0.62, 0.37    (State: S=sleeping R=running, W=waiting)

  PID USER     STATUS   RSS  PPID %CPU %MEM COMMAND
26661 admin    S        26M     1 77.0 10.5 python2.6
24669 admin    S        41M     1  0.4 16.6 python
26532 admin    S       3824  2985  0.3  1.4 smbd
26806 admin    R        920 26711  0.3  0.3 top
 6737 admin    S       2724     1  0.1  1.0 upnpcd
26351 admin    SW         0     2  0.1  0.0 flush-8:0
 3003 admin    S <     7332     1  0.0  2.8 apache
 6734 httpdusr S <     4864  3003  0.0  1.9 apache
 6735 httpdusr S <     4852  3003  0.0  1.8 apache
26547 admin    S       3472  2985  0.0  1.3 smbd

This is while SABNZBD is running and downloading at 1.2 MB/s (max for this connection). This is after a repair and restart.

I've been checking some things, and i've noticed when adding a NZB this problem occurs when the nzb still says "checking" and when it's still on checking it can't download. And i've disabled this option "Check before download". And this seemed to solve the problem. Only thing that still bothers me, these nzbs get added during the day when SANBZBD is on pause, so the nas should have enough time to check the download. For now i've disabled this option.