Failing repairs on content not identified as corrupted

[nzbtuxnews]

Hello,

there is some content I am trying to download from a certain category (lets mention a*ias) from nzbmatrix,  and so far 4 attempts have failed to complete due to errors like

" Repair failed, not enough repair blocks (XX short)"

I've checked on my indexing site (nzbmatrix) and nobody reported the content to be faulty or corrupted.

Anybody can provide assistance to solve this issue?  Frustrating because since a few weeks ago, evething I download from sab seems to be corrupted.  Pointless if you ask!
Thanks!

[shypike]

Which release do you use?
Do the NZB files look complete?
(Click on the job title to see the underlying files).
Do you see messages about missing articles or other warnings?

[nzbtuxnews]

I am still using 0.6.0RC1 but tried to upgrade to 0.6.1 and unfortunately I lost my queue during the upgrade.  I reverted to 0.6.0RC1 but do you think upgrading to 0.6.1 would help in any way?

***

[shypike]

I think the loss of the queue was just an unfortunate accident.
The difference between 0.6.1RC1 and 0.6.1 is only an attempt
to fix problems with the NZB file extension coupled to SABnzbd.

BTW: did you de-install 0.6.1RC1 before upgrading and did you
tell it to remove data too?

[nzbtuxnews]

hey shypike,

I can restore the queue right?  The files are all still in the incomplete folder.  Just gone from the SAB interface.

What I did to "upgrade" is the following:

1-I shutdown SAB 0.6.0RC1
2-I renamed the folder so it does not interfere with 0.6.1
3-I unpacked 0.6.1 in my apache folder
4-I copied the sabnzbd.ini over the new install
5-I started SAB normally.

All went fine, except my queue was gone!

I am not sure how to uninstall the old version prior to installing a new version since its only a matter of moving/renaming/deleting the folder right??

No I did not tell to delete anything.

One thing I am convinced most people would appreciate is some kind of automatic update feature like Sickbeard or Couchpotato.  Whenever a new version is released (or updates from the GIT), we can update without fiddling with the real files/folders.  Not a big deal if you ask because the way you guys built SAB is already simple enough...  I'd rather have such functions to update/maintain instead of other functions like bandwidth management or data plotting like I've seen on some threads... --> keep it simple but efficient.

Thanks!

[nzbtuxnews]

OK so I've upgraded to 0.6.1 and tried to re-import my queue, assuming this can be done.

Here's what I did:

1- Deleted all __ADMIN__ folders in my job folders (all located under /incomplete)
2- Went back to my indexing site to retrieve the NZB file again
3- At that moment, SAB's queue was completely empty, then I added the jobs using the "Add NZB" button at the top of the window.
4-  SAB started the job from scratch (*from zero) and was not reusing the existing files in /incomplete

I also noticed that SAB creates a Folder.Bla.Bla.Bla.1 in my incomplete folder.

The NZB File & the job folder are exactly name the same way...

WHat to do next?

[shypike]

4-I copied the sabnzbd.ini over the new install

That's where it went wrong.
SABnzbd by default places it admin folder where the sabnzbd.ini file is.
So assuming you had the INI file in SABnzbd's program folder,
it's now gone to the renamed RC1 folder.

You could have picked up the "orphaned" jobs on the Status page.
By deleting admin stuff in "incomplete" you messed up things with unpredictable results.

[nzbtuxnews]

Yeah I think I'll wait until my queue is complete and then upgrade clean....

Thanks for the indications!

[nzbtuxnews]

I'm back here because I am pretty sure I know what happened...  You see my main usenet provider is supernews (giganews) and have a retention of 800 days.

Yesterday I tried to download something else that had been posted over 950 days ago.  It failed.  Posting on the forums of my indexing site the people suggested that anything I would try with over 800 days on supernews was likely to fail and I should try a different usenet server such as astraweb.

So let's say I purchase an account with astraweb, add their server to sab and configure it, how can I use this new server to retry the failed file?  Precisely if Sab says that there is 50 missing blocks, how can I tell Sab to use my new server to retrieve these missing blocks instead of using my base server (supernews) or even my block server??

Is Sab smart enough to try all servers in priority before stating that there is missing blocks?

[shypike]

When missing articles are scattered over multiple files, this is a problem.
SABnzbd will not retry individual articles.
Find the most damaged file(s), remove them and retry.
Those file(s) will be fully downloaded with the retry.
SABnzbd will just try all servers again, first the primaries and then the backups.

[nzbtuxnews]

When the job is in the history queue, howcan you see which files are damaged or incomplete?  Clicking on the job name will only show a small information window not the list of Rar files like the jobs in the download queue...

Also a side question:  one of my jobs was very large and to prevent using too much bandwidth I paused the job while I kept Sab running.  today I had a power outage so my server rebooted and the job got resumed resulting in completion and over bandwidth for this month by over 50gb... Will probably cost me 50$ extra on top of my normal bill...  How do you prevent such things?

[shypike]

Sorry about your overage. Not nice being screwed over by your ISP.

If you pause jobs individually, that state will be preserved.
Although a power cut will prevent the state from being saved to disk.
That explains why it resumed after a restart.

It's also possible to start SABnzbd in paused state:
Start it with the -p command line option (you'll need to modify the shortcut or the script that you start it with).

I'll see if I can improve matters by saving a job's state as soon as you pause it.

[nzbtuxnews]

That's what happens when you live in Canada.. Weak government that do not regulate and abusive isp's...

Anyways what about the damaged files?  Like I said I don't know how to identify the damaged files... Do I simply unrar and see which Rar archive unrar barf on?

[shypike]

Damaged usually have a size that's not a round number.
50M is a popular number.
If bandwidth is so expensive, I suggest you use QuickPar to
assess what the damage is and get as much par2 files as possible.
The most damaged files you could re-download again.
Use the "Download-only" option of SABnzbd to get the files.
So you'll need to do the job in the old-fashioned manual way for once.

[nzbtuxnews]

So does quickpar is available for Linux systems?

Also I'll have to perform this procedure at least 6 times because I've got a long list if failed jobs in my history queue...

Is there a tutorial somewhere so I can be guided step by step for the first time?