Large queue nzb's or large nzb's makes download speed 'very' slow

[pikkedonker]

HI,

I own a synolgy DS409+ (Freescale MPC8533 / Intern mem 512 Mb) and I've noticed that when I chcuk my queue full with nzb's or a very large nzb file my speed goes intermittently down/up. The longer the queue, the lower the DL speed. According 2 the release notes this problem (or behaviour) should B solved in 0.5.0. (this is my standard install now). But its still a symptom. I tried also 0.5.1.RC2. Still the same problem.

In another thread I've red that a backup server maybe involved. (How I dont know precise...just a noob) But my NAS isn't running 1. I've check several times what services r up, and not ever having used the backup facillity of the synolgy. I'm pretty sure it's off.

Any1 have any idea?

Greetz

[shypike]

Are you using a backup (fill) server?
If so that may be the cause of the problem.
Recently discovered and will be solved in 0.5.2

Check if the "cache" folder consists of a number of SABnzbd_nzo_* files
sized below 100K and lots of SABnzbd_nzf_* files.
If there are just a few nzf files and the nzo files are large (>500K),
then you've been hit by this bug.

[pikkedonker]

Thx for the quick response! Nope I'm not using any backup server of any kind. But I've seen an AWEFULL lot of files in the cach when this happens (dont know about the size and if it's those nzo files :-( ). I've got my SABNZBd up and running again (remember noob here!) through a clean install and winSCP intervention. I dare not try it again tbh. For now i'll download 1 nzb at a time and that works perfectly! I'll wait for 0.5.2 and test it again! But thx for now!! Will report back! ;-)

[shypike]

With "backup server" I mean a Usenet server that is flagged as "backup"
in Config->Servers. It has nothing to do with the backup of the NAS.

[pikkedonker]

AHA! As I was saying....NOOB HERE! Thx! I have indeed 2 backup servers configured! Will try that this afternoon when I get home! I'll report back! Thx for ur patience!

Greetz

[shypike]

The problem will not go away immediately.
Only when the queue has become empty, will new jobs start behaving properly.
i.e. if you remove the backup servers.

[pikkedonker]

Yup queue was empty. Removed the backup newsservers , restarted the NAS (and the service to B on the save side). Loaded the really big nzb that gave me problems and presto! Downloading like a bat outta hell!

Thx so much shypike!
Really looking forward to 0.5.2. Think u guys R doing a hell of a job!

Greetz!

[pikkedonker]

My download is nearing the end and now the problem returns (I'd say it started at 90% of the downlaod). I checked the cache Dir. It's not nearly as full as it was before the removal of the backup servers, but there is a NZO file near the end of 37 MB. Will removing that NZO file clear me of the bug?

Greetz

[shypike]

A 37M NZO file is very suspicious.
Last time I tested, an NZB for downloading 20G gave an NZO of less then 2M.
The 37M NZO will lockup the queue on every queue save,
which can be quite frequent.
How big is the download associated with this NZO?

[pikkedonker]

It's about 120-140 G! I know it's a lot! (maybe it's pushing it) I also checked cpu (50-60% load) and memory (never above 60%)
The queue and the program still run however!! And like u said at some point my download is max and with every queue save? it drops to 5-10% of normal speed.

Greetz and thx again for ur time

[shypike]

I assume this large NZO is still the victim of the backup-server bug.
I do advise that when you have such large items to set
the option "Only get articles for top of the queue" to "on" in Config->Switches.

[pikkedonker]

Will do! Thx for the reply and got another LARGE nzb lined up. So i'll report back in a day or 2 ;-) .
Will adjust the options in config. (Just checked and the option was already on!) So I'll dail it down for now and wait for 0.5.2 ;-)

Greetz and thx for ur time

[pikkedonker]

Hmm, he's verifying now and these r the errormessage I get:

ERROR:
[15/Apr/2010:22:20:10]  Traceback (most recent call last):
  File "/volume1/@appstore/sab2/SABnzbd-0.5.0/cherrypy/_cpwsgi.py", line 79, in setapp
    s, h, b = self.get_response()
  File "/volume1/@appstore/sab2/SABnzbd-0.5.0/cherrypy/_cpwsgi.py", line 219, in get_response
    response = self.request.run(meth, path, qs, rproto, headers, rfile)
  File "/volume1/@appstore/sab2/SABnzbd-0.5.0/cherrypy/_cprequest.py", line 579, in run
    raise cherrypy.TimeoutError()
TimeoutError

The queue has no NZO file and still something is hoggin my disks verifying is also slow.

Greetz

[shypike]

Start once with the --clean option to clear out any corrupted data.

[pikkedonker]

I've called it tbh.
Something is hoggin the sabnzb cannot get into the webserver anymore and that restart -clean is on a Synology without a masters in space-engeneering not 2 B found, I'm afraid. (tried WinSCP and Telnet)
Think the lesson for now (and for me, the noob!) is to just stop the service. remove app en clean install. Then just download small nzb packages and wait for 0.5.2. ;-)
Good luck with 0.5.2. make us proud! ;-)

And thx for ur patience m8!

Greetz Eric