'NoneType' object has no attribute 'remove' [0.7.0B3]

[hbz]

EDIT- Sorry if I have posted this in the wrong forum as I haven't proved it's a bug with the recent beta release.

This could be related to the delete/don't download samples switch, but I disabled it and I still get it. It happens with almost every nzb now. I completely purged my install and started fresh again...and for some reason this problem just returned. sab is run as root so there shouldn't be any permissioning problems.

Code: Select all

2012-05-10 08:09:45,580::INFO::[postproc:559] Par2 check finished on NZB.I.Downloaded-WITHPARS
2012-05-10 08:09:45,582::INFO::[notifier:87] Sending registration to localhost:23053
2012-05-10 08:09:46,086::INFO::[postproc:504] Cleaning up NZB.I.Downloaded-WITHPARS (keep_basic=True)
2012-05-10 08:09:46,090::INFO::[misc:1230] Cannot remove folder /tmp/NZB.I.Downloaded-WITHPARS.1/__ADMIN__
2012-05-10 08:09:46,090::INFO::[misc:1230] Cannot remove folder /tmp/NZB.I.Downloaded-WITHPARS.1/__ADMIN__
2012-05-10 08:09:46,121::ERROR::[postproc:127] Failed to remove nzo from postproc queue (id) SABnzbd_nzo_8ze0xg
2012-05-10 08:09:46,121::INFO::[postproc:84] Saving postproc queue
2012-05-10 08:09:46,122::INFO::[__init__:877] Saving data for postproc1.sab in /root/.sabnzbd/admin/postproc1.sab
2012-05-10 08:09:46,122::INFO::[downloader:243] Post-processing finished, resuming download
2012-05-10 08:09:46,124::INFO::[postproc:273] Starting PostProcessing on NZB.I.Downloaded-WITHPARS => Repair:True, Unpack:True, Delete:True, Script:None, Cat:*
2012-05-10 08:09:46,125::INFO::[notifier:87] Sending registration to localhost:23053
2012-05-10 08:09:46,625::INFO::[postproc:530] Par2 check starting on NZB.I.Downloaded-WITHPARS
2012-05-10 08:09:46,626::INFO::[postproc:545] Running repair on set NZB.I.Downloaded-WITHPARS
2012-05-10 08:09:46,631::INFO::[newsunpack:1292] Quick-check of file NZB.I.Downloaded-WITHPARS.r11 failed!
2012-05-10 08:09:46,631::INFO::[newsunpack:787] Scanning "/tmp/NZB.I.Downloaded-WITHPARS.1/NZB.I.Downloaded-WITHPARS.vol00+01.par2"
2012-05-10 08:10:40,248::INFO::[newsunpack:969] Verified in 53 seconds, repair is required
2012-05-10 08:10:40,249::INFO::[newsunpack:1007] Need to fetch 27 more blocks, checking blocks
2012-05-10 08:10:40,250::INFO::[newsunpack:1029] 8 blocks available
2012-05-10 08:10:40,250::ERROR::[newsunpack:814] Error 'NoneType' object has no attribute 'remove' while running par2_repair on set NZB.I.Downloaded-WITHPARS
2012-05-10 08:10:40,251::INFO::[newsunpack:815] Traceback:
Traceback (most recent call last):
  File "/usr/share/sabnzbdplus/sabnzbd/newsunpack.py", line 792, in par2_repair
    setname, joinables)
  File "/usr/share/sabnzbdplus/sabnzbd/newsunpack.py", line 1053, in PAR_Verify
    nzo.add_parfile(extrapar_list.pop())
  File "/usr/share/sabnzbdplus/sabnzbd/nzbstuff.py", line 1011, in add_parfile
    parfile.extrapars.remove(parfile)
AttributeError: 'NoneType' object has no attribute 'remove'

[hbz]

I think this actually is a problem with the yENC decoder or something, because I remember setting my client only to download (not repair/delete) and it was missing 1 block from every file. It's not a problem with my provider or the age either. I'll have to run some more tests.

[shypike]

yEnc has no relation with par2 verification and repair.
yEnc decoding always needs to be done even for download-only.
It could be that yEnc records are so damaged that bad par2 files get created.
There could be a bug in SABnzbd that is triggered by posts damaged in a specific way.
I would need to have the NZB file to examine this problem.
Email it to bugs@sabnzbd.org and include your Usenet provider's name and the URL of this post.

[hbz]

Yeah I keep examining logs where this happens and it could be my provider. I'm sorry to have misposted this. Feel free to delete!

[hbz]

I feel like this is related: https://forums.sabnzbd.org/viewtopic.php?f=16&t=10667

[shypike]

Possibly, but not likely because that issue results in server disconnects as well.
It's a bug any how because exceptions like you see in your log should not occur.
Anyway, I just published Beta5 which solves the other problem.
I doubt whether it will solve yours, but please let me know.

[hbz]

Thanks so much for your help here and I'll be sure to report back.

[i814u2]

Any update on this? I'm having the same issue. I've had it since roughly Beta5 (not sure if it's the data or the software though). I've upgraded to Beta6 and Beta8 with no change.

Please let me know what information I can provide that will help.

newshosting as the provider, not all downloads experience this (just some that are having issues with missing/bad blocks and needing repair. They are only a few months old but apparently have some damaged blocks).

[shypike]

Please list the exact logging.
People assume too easily that they have "the same problem".
It cannot be the same problem because this specific bug was solved.
That could still mean that another bug is hitting you, but I cannot tell without more info.

[i814u2]

Understood. The errors looked similar to me, which is why I made that assumption.
Here is an example from the log, to show what I'm seeing:

Code: Select all

2012-05-29 11:58:31,040::INFO::[postproc:280] Starting PostProcessing on boardwalk ctrlhd 720p => Repair:True, Unpack:True, Delete:True, Script:None, Cat:tv
2012-05-29 11:58:31,040::INFO::[postproc:544] Par2 check starting on boardwalk ctrlhd 720p
2012-05-29 11:58:31,040::INFO::[postproc:559] Running repair on set Boardwalk.Empire.S02E07.720p.HDTV.DD5.1.x264-CtrlHD
2012-05-29 11:58:31,069::INFO::[newsunpack:1292] Quick-check of file Boardwalk.Empire.S02E07.720p.HDTV.DD5.1.x264-CtrlHD.part33.rar failed!
2012-05-29 11:58:31,069::INFO::[newsunpack:787] Scanning "/mnt/tank01/data/NZB-InProgress/boardwalk ctrlhd 720p/Boardwalk.Empire.S02E07.720p.HDTV.DD5.1.x264-CtrlHD.vol000+01.par2"
2012-05-29 11:58:45,640::INFO::[newsunpack:969] Verified in 14 seconds, repair is required
2012-05-29 11:58:45,640::INFO::[newsunpack:1007] Need to fetch 33 more blocks, checking blocks
2012-05-29 11:58:45,640::INFO::[newsunpack:1029] 182 blocks available
2012-05-29 11:58:45,640::ERROR::[newsunpack:814] Error 'NoneType' object has no attribute 'remove' while running par2_repair on set Boardwalk.Empire.S02E07.720p.HDTV.DD5.1.x264-CtrlHD
2012-05-29 11:58:45,641::INFO::[newsunpack:815] Traceback: 
Traceback (most recent call last):
  File "/usr/share/sabnzbdplus/sabnzbd/newsunpack.py", line 792, in par2_repair
    setname, joinables)
  File "/usr/share/sabnzbdplus/sabnzbd/newsunpack.py", line 1053, in PAR_Verify
    nzo.add_parfile(extrapar_list.pop())
  File "/usr/share/sabnzbdplus/sabnzbd/nzbstuff.py", line 1011, in add_parfile
    parfile.extrapars.remove(parfile)
AttributeError: 'NoneType' object has no attribute 'remove'

This particular nzb was manually "created" via binsearch.info and then manually added to SAB.
If I'm completely missing something, or need to provide you with more info, just let me know.

Thanks for the prompt attention, definitely appreciated.

[i814u2]

I also noticed that as I've been updating, between beta releases, that I receive this error:

Code: Select all

2012-05-29 12:20:03,676::ERROR::[__init__:853] Loading /home/<username_redacted>/.sabnzbd/admin/future/SABnzbd_attrib failed

I'm not sure if that's my fault, something unrelated, or something related to this issue. I'm just mentioning it in case it helps you out.

[shypike]

This particular nzb was manually "created" via binsearch.info and then manually added to SAB.
If I'm completely missing something, or need to provide you with more info, just let me know.

It's a bug all right, but one completely unrelated.
Can you email the NZB file to bugs@sabnzbd.org ?
I know where to look in the code, but having an NZB that triggers the bug is so much easier
Thanks for reporting.

[shypike]

I also noticed that as I've been updating, between beta releases, that I receive this error:

Code: Select all

2012-05-29 12:20:03,676::ERROR::[__init__:853] Loading /home/<username_redacted>/.sabnzbd/admin/future/SABnzbd_attrib failed

I'm not sure if that's my fault, something unrelated, or something related to this issue. I'm just mentioning it in case it helps you out.

Did you lose any data?
It looks like this is either an incompletely removed entry where SABnzbd gets NZBs itself from a site.
or a damaged entry of such an attempt.
Could be a bug or the result of a bluntly terminated SABnzbd session.

[i814u2]

I haven't lost any data that I've been downloading. And haven't noticed any abrupt stop of SAB at all. I either manually stop/start the service if it's hung, or I use the "Repair" option for the queue when a few items start collecting (when it attempts to download more chunks, it frequently will double-up on the queue entries (I've even had a few for a single download).
I suppose it could have happened either during upgrade, or simply when using the repair feature. Sounds like it isn't related though, so I'll keep an eye on it and start a new post if I can provide any concrete information.

At one point, I manually removed the admin folder to reset (lost all the history, of course) and things were fine then. But it's happened again, which is why I mentioned it. However, like I mentioned, I both upgraded AND had download issues (used repair option once I had too many items apparently doing the same task). So I can't say which one appears to be the cause.

[shypike]

It's a bug all right, but one completely unrelated.
Can you email the NZB file to bugs@sabnzbd.org ?
I know where to look in the code, but having an NZB that triggers the bug is so much easier
Thanks for reporting.

Can you please email the NZB?