I'm experience a few big issues since the latest update on the development branch.
First, the settings pages will no longer accept any config changes, just shows a popup saying an error occurred. Though some setting do end up sticking when I revisit the page, but others do not, for example, on the 'Web Interface' page, the NZb search interval saves, but the comic location does not.
Looking at other posts, I see you mention an issue with the sabnzbd check since you changed from the full to the nzb key, so I would take a guess that setting the page saves before the sab check are good, and then it fails on the sab check and doesn't apply the remaining settings.
The next issue is a big one, when searching for comics, mylar has always inserted a delay between searches, I've always had this set as 1 minute and looking back through my logs I can see that in always worked, stating 'pausing for 60 seconds before continuing to avoid hammering', however, since the latest update my logs now show 'pausing for 111111111111111111111111111111111111111111111111111111111111 seconds before continuing to avoid hammering'.
This ultimately mean it will search for one comic and then pauses it self for, well a very very long time, none of us will even be remembered by the time it finishes.
I re installed Mylar into a new directory and onto a new system and i get the settings issue on both, but i couldn't test the pause delay on that as the settings issue meant i could setup Mylar to allow me to add comics.
This is on Windows Server 2008 R2 with Python 2.7.2
Issues since last update
-
rotox
- Posts: 13
- Joined: Mon Feb 24, 2014 3:07 pm
Re: Issues since last update
I had the same problem a while back (http://mylar.forumatic.com/viewtopic.php?f=8&t=119#p389). The only way that I was able to resolve it was to disable the RSS searches and just use NewzNab. The problem with doing that is as follows
(http://mylar.forumatic.com/viewtopic.php?f=8&t=119#p503). So as long as you don't use torrents you'll be fine.Well the RSS Feed option is a global option (I perhaps should elude to that in some way on the main page). Disabling it will turn off the RSS Feed Indexing for ALL indexers (even torrents). - evilhero
-
j0rdan
- Posts: 3
- Joined: Sat Apr 05, 2014 11:44 am
Re: Issues since last update
OK, so, RSS was already disabled.
I turned off Experimental search (had to shut down and manually edit the config, as the web interface wouldnt take the setting change.
After restart I force a check for all wanted comics and it now sleeps for 60 seconds between searches, BUT, it sleeps 3 time between each search. log below.
That is until I make (what appears to be) any change in the settings, where it instantly makes the next line in the log 'pausing for 11111111111111111111111111111111111111111111111111111 seconds' (though this appears to be whatever number entered into the 'Search delay' option, for eg. it you put 2 in there it will say pausing for 222222222222.... seconds, if you put 0 it says pausing for 000000...... second etc) and it then errors out.
logs below shows the triple pausing on each search, I then, between 15:21 and 15:22 make any change in the setting page and the next line is the long pause followed by an error.
The error doesn't crash Mylar, and isn't shown in the log, only in the python window, the web interface continues to respond, I can even force a new search from the web interface, although this just brings up the same error lines in the python windows.
So from the error I would guess the delay isn't being stored in the correct format?, although the search delay isn't the value that was changed to get this error, the error appears on any config change what so ever.
ps. sorry I didn't include the python output in my first post, I run Mylar as a service so don't normally see it so usually just really on the logs.
I turned off Experimental search (had to shut down and manually edit the config, as the web interface wouldnt take the setting change.
After restart I force a check for all wanted comics and it now sleeps for 60 seconds between searches, BUT, it sleeps 3 time between each search. log below.
Code: Select all
2014-04-05 15:02:13 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 15:01:12 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 15:00:12 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 15:00:12 INFO Shhh be very quiet...I'm looking for Clive Barker's Hellraiser: The Road Below issue: 3 (2012) using Usenet-Crawler (newznab)
2014-04-05 14:59:11 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 14:58:10 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 14:57:09 INFO pausing for 60 seconds before continuing to avoid hammering
2014-04-05 14:57:09 INFO Shhh be very quiet...I'm looking for Clive Barker's Hellraiser: The Road Below issue: 3 (2012) using https://nzbs.in (newznab)logs below shows the triple pausing on each search, I then, between 15:21 and 15:22 make any change in the setting page and the next line is the long pause followed by an error.
Code: Select all
05-Apr-2014 15:15:01 - INFO :: Thread-13 : Shhh be very quiet...I'm looking for Clive Barker's Hellraiser
: The Road Below issue: 3 (2012) using https://nzbs.in (newznab)
05-Apr-2014 15:15:01 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:16:02 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:17:03 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:18:04 - INFO :: Thread-13 : Shhh be very quiet...I'm looking for Clive Barker's Hellraiser
: The Road Below issue: 3 (2012) using Usenet-Crawler (newznab)
05-Apr-2014 15:18:04 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:19:05 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:20:05 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:21:05 - INFO :: Thread-13 : Shhh be very quiet...I'm looking for Clive Barker's Hellraiser
: The Road Below issue: 3 (2012) using https://01100111011001010110010101101011.info/ (newznab)
05-Apr-2014 15:21:05 - INFO :: Thread-13 : pausing for 60 seconds before continuing to avoid hammering
05-Apr-2014 15:22:07 - INFO :: Thread-13 : pausing for 11111111111111111111111111111111111111111111111111
1111111111 seconds before continuing to avoid hammering
Exception in thread Thread-13:
Traceback (most recent call last):
File "C:\PYTHON27\LIB\threading.py", line 552, in __bootstrap_inner
self.run()
File "C:\PYTHON27\LIB\threading.py", line 505, in run
self.__target(*self.__args, **self.__kwargs)
File "C:\Apps\Mylar\mylar\mylar\search.py", line 1505, in searchforissue
foundNZB, prov = search_init(comic['ComicName'], result['Issue_Number'], str(ComicYear), comic['Comic
Year'], Publisher, IssueDate, StoreDate, result['IssueID'], AlternateSearch, UseFuzzy, ComicVersion, SARC
=None, IssueArcID=None, mode=mode, rsscheck=rsscheck, ComicID=result['ComicID'])
File "C:\Apps\Mylar\mylar\mylar\search.py", line 294, in search_init
findit = NZB_SEARCH(ComicName, IssueNumber, ComicYear, SeriesYear, Publisher, IssueDate, StoreDate, n
zbprov, nzbpr, IssDateFix, IssueID, UseFuzzy, newznab_host, ComicVersion, SARC=SARC, IssueArcID=IssueArcI
D, ComicID=ComicID)
File "C:\Apps\Mylar\mylar\mylar\search.py", line 645, in NZB_SEARCH
time.sleep(pause_the_search)
TypeError: a float is requiredSo from the error I would guess the delay isn't being stored in the correct format?, although the search delay isn't the value that was changed to get this error, the error appears on any config change what so ever.
ps. sorry I didn't include the python output in my first post, I run Mylar as a service so don't normally see it so usually just really on the logs.
-
evilhero
- Site Admin
- Posts: 2887
- Joined: Sat Apr 20, 2013 3:43 pm
- Contact:
Re: Issues since last update
Well there's a solution for one, and the other I have to remove something from the last update in order for things to work again.
Whenever you make any configuration change, you need to restart Mylar. Not doing so causes Mylar to reuse the values for things that may have changed and really wonky things start happening that can't be explained or duplicated for the most part. I'll look into if this is because of not restarting, but won't be able to test things until later today most likely (unless you can confirm that yourself before I can).
The pausing three times before continuing searching is normal, Mylar will pause X seconds between search iterations (1,01,001) so if the issue is from 1-10,it would search up to three times.. Issue 10-99 would search up to two times.. And 100+ would search only once. This is because searching for the number variations gives the most possible results and is necessary because searching for just '1' will give different results that '01' and even '001'.
The config is actually die to the test Sabnzbd button in the last update.. I'm going to remove that option entirely for the time being until I can work out a better way to do things, and in doing so the configuration will be able to be saved as was normal just before the last update (my apologies for letting that one slip).
As an aside I have also noticed discrepancies between various browsers when accessing the same pages - in Firefox, the auto-populate sabnzbd api key doesn't work, but in ie it works fine. I'll be looking into this more over the course of the next few days in an attempt to get better cross browser performance.
Whenever you make any configuration change, you need to restart Mylar. Not doing so causes Mylar to reuse the values for things that may have changed and really wonky things start happening that can't be explained or duplicated for the most part. I'll look into if this is because of not restarting, but won't be able to test things until later today most likely (unless you can confirm that yourself before I can).
The pausing three times before continuing searching is normal, Mylar will pause X seconds between search iterations (1,01,001) so if the issue is from 1-10,it would search up to three times.. Issue 10-99 would search up to two times.. And 100+ would search only once. This is because searching for the number variations gives the most possible results and is necessary because searching for just '1' will give different results that '01' and even '001'.
The config is actually die to the test Sabnzbd button in the last update.. I'm going to remove that option entirely for the time being until I can work out a better way to do things, and in doing so the configuration will be able to be saved as was normal just before the last update (my apologies for letting that one slip).
As an aside I have also noticed discrepancies between various browsers when accessing the same pages - in Firefox, the auto-populate sabnzbd api key doesn't work, but in ie it works fine. I'll be looking into this more over the course of the next few days in an attempt to get better cross browser performance.