Newznab - is it even searching?

[jifka]

I have a newznab instance set up on localhost, and it's configured to only pull from 1 group, a.b.comics.dcp. I've loaded it with 1 year's worth of data, and I can see the binaries listed in the Raw Search in newznab, but nothing comes up in Releases.

In Mylar I've set up the newznab to point to localhost (with trailing slash) and API key. I've added comics and marked issues as wanted and refreshed. All I get is this:

14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Updating complete for: Saga
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Now checking files for Saga (2012) in /Users/xxxxxx/comics/mylar/Saga-(2012)
14-Mar-2013 23:20:51 - DEBUG   :: Thread-13 : comic: Saga
14-Mar-2013 23:20:51 - DEBUG   :: Thread-13 : Looking in: /Users/xxxxxx/comics/mylar/Saga-(2012)
14-Mar-2013 23:20:51 - DEBUG   :: Thread-13 : you have a total of 0 Saga comics
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : I've found 0 / 10 issues.
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Checking this week's pullist for new issues of Saga
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Checking the Weekly Releases list for comics I'm watching...
14-Mar-2013 23:20:51 - DEBUG   :: Thread-13 : You are watching for: 1 comics
14-Mar-2013 23:20:51 - DEBUG   :: Thread-13 : There are 0 comics this week to get!
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Finished checking for comics on my watchlist.
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Attempting to grab wanted issues for : Saga
14-Mar-2013 23:20:51 - INFO    :: Thread-13 : Finished grabbing what I could.

I'm definitely not ruling out the possibility (probability?) that my newznab configuration is not correct and is not providing results, but I don't see anything in the Mylar log to indicate that the newznab server is even being hit.

Here's a screenshot of the Mylar setup:

And here's a shot of the newznab raw search showing content that should match:

[evilhero]

I'm betting you didn't install Sphinx when you setup the newznab. For whatever reason, you need to have it enabled in order for anything to get found ( It was happening to myself and installing Sphinx solved it instantly )

[jifka]

I'll have to work on that. I'm on OS X Lion, and installing Sphinx is...um...non-trivial.

[evilhero]

Are you running the developmental or master build? The only other thing it could be is that you have a blank entry for a newznab. Open up the config.ini and scroll down until you get to the newznab section and paste it here (remove the keys first thi)

[synapseattack]

Hey, not sure why you think sphinx needs to be installed. I have mine running fine without it. So far it finds just about everything I throw at it. I'm running the latest development build of Mylar with newznab installed on the same system.

[jifka]

I tried running the dev build a while ago and got all sorts of errors, so I went back to the master build.

@synapseattack, would you mind sharing your newznab configuration? Do you have some custom regex? Which newsgroup(s) are you hitting?

[evilhero]

synapseattack: The reason why is because it's a text based search engine that is extremely good at what it does in a much more efficient way than the stock. If you're using the cusom regex's from the plus version, installing sphinx will just ensure that you don't miss something - you don't need to by all means, I'm just saying what had solved my problem was installing sphinx even when I have the plus version.

The main reason you're not matching is probably because the releases aren't getting regex'd as they should...and all the crap at the beginning of those files you had posted (Novus distribution, 0-day, zero-day, etc) gets removed when you use the proper regex' and then you'll get your matches as Mylar assumes if it's not exactly what it's looking for it no-matches.

So um, pay for the plus version and support the program - and as an added bonus, you'll get matches :) (not saying that's the way it has to be done, but it is about supporting the programs you use).

[synapseattack]

Oh, in that case I am using a custom regex as I use the Plus version. If you want to shot me a email to something I'll be happy to send you the regex stuff for the comic groups. But I agree with @evilhero, if you can afford it please support newznab by getting the plus subscription. It's really worth it.

[jifka]

Yessir, I bought the plus version, but didn't know what the deal was and why nothing was showing up. I'll send a direct message to get your regex's, but still have this lingering doubt...

I still don't know if Mylar is even hitting my newznab install. It doesn't seem to indicate a failure, but the logs (AFAIK) aren't showing an attempt to hit it.

[jifka]

OK, I think I have this figured out, finally. It was partially my fault but partially difficult to debug because I wasn't getting much feedback. Here's the end result:

1 - My newznab+ configuration was not complete. Because I had failed to enter the correct serial number (yes I purchased, but chopped off a character in the copy/paste), I never got the updated regex listings for alt.binaries.comics.dcp. Once that cleared up, I was able to backfill and generate releases. Chalk that up to MY failure.

2 - Once that was fixed, I configured the connection from Mylar to newznab and BOOM I see errors in the Mylar log saying it can't send nzbs to SABnzbd. Ooops - I forgot to spin that up. But once I did, the errors went away, but nothing in the Mylar log indicated any files getting sent to SAB. And SAB wasn't reporting anything inbound. Again, my failure since I copied the wrong API key. SAB provides 2 keys, one is called the API key and the other is the NZB key - I needed to use the NZB key. So, there's some confusion there. Not only because the Mylar interface says "SABnzbd API:" but because Mylar didn't indicate a failure. I can't tell who's at fault for the latter - is SAB not reporting a failure back to Mylar? Is Mylar not logging the failure?

Right now Mylar is working, searching newznab, sending NZBs to SAB, which downloads correctly...but then it breaks down again. The .cbr/.cbz files are there in the SAB download path and I have that path configured in Mylar under "SABnzbd Download Directory", and I have post-processing enabled. But I'm not seeing Mylar locate the files to put them in the proper directory. Isn't Mylar supposed to do that? The comic folders are created by Mylar, tho.

Manually refreshing / rechecking files doesn't seem to help, either.

[evilhero]

ok so first thing, switch to the developmental build. The master build gets updated maybe once every 2 months , so if something happens in regards to bugs it will be resolved in that build.

Second, it IS the API key, not the NZB key..I'm not sure why yours is different, but you need to feed in the api key

The last bit in regards to post-processing. ..did you edit the autoProcessComics.cfg.sample file located in mylar root/sabnzbd to coincide with your mylar settings ( and resave it without the .sample at the end ) and then move the .cfg, autoProcessComics.py and ComicRN.py into your SABbzbd scripts directory (in the readme.md file). Create a category in sab identical to the same stated in mylar, and make sure you have sab pointing to the same dest dir you put in mylar for downloads (you don't need to specify a dl dir for sab in mylar, its optional and the downloaded directories are deleted after a successful move).

[jifka]

I had been on the development branch, so that's not an issue...and I switched it to the API key and it works the same, so I must've had a bad test somewhere in there.

As for the SAB configuration for post-processing - yep, that did the trick. I think there are so many moving parts, that if you're not careful (and miss something in the readme - duh!) then it's tough to debug.

Keep up the good work!