14 replies to this topic

[socks]

Hi,

 

I'm  running latest OpenPLi4 on two ET10000 boxes. On both boxes enigma2 is crashing during the night after EPG refresh has finished.

EPG refresh is scheduled every night between 4 and 7am.

I'm using a FAT32 formatted USB stick labeled as "ET10000" to save the epg.dat file to.

 

This is my EPG refresh plugin configuration:

config.plugins.epgrefresh.begin=4:0                                                                
config.plugins.epgrefresh.interval=2                                                               
config.plugins.epgrefresh.save_epg=true                                                            
config.plugins.epgrefresh.start_on_mainmenu=true                                                   
config.plugins.epgrefresh.adapter=record                                                           
config.plugins.epgrefresh.lastscan=1435459712                                                      
config.plugins.epgrefresh.enabled=true                                                             
config.plugins.epgrefresh.delay_standby=10                                                         
config.plugins.epgrefresh.show_help=false                                                          
config.plugins.epgrefresh.interval_seconds=60                                                      
config.plugins.epgrefresh.add_to_refresh=0                                                         
config.plugins.epgrefresh.end=7:0                                                                  
config.plugins.epgrefresh.force=true
config.plugins.epgrefresh_extra.timeout_autozap=90                                                 
config.plugins.epgrefresh_extra.show_autozap=true                                                  
config.plugins.epgrefresh_extra.epgcachepath=/media/ET10000/                                       
config.plugins.epgrefresh_extra.cachesavetimer=12                                                  
config.plugins.epgrefresh_extra.save_backup=true
config.misc.epgcache_filename=/media/ET10000/epg.dat

This are the last few lines of the enigma2 crash log:

<578946.952943> [eEPGCache] abort non avail schedule other reading            
<578946.953588> [eEPGCache] abort non avail netmed schedule reading           
<578946.954208> [eEPGCache] abort non avail netmed schedule other reading     
<578946.954922> [eEPGCache] abort non avail FreeSat schedule_other reading    
<578946.955625> [eEPGCache] abort non avail viasat reading                                          
<578954.949047> [eEPGCache] nownext finished(1435632770)                  
[EPGRefresh] Maybe zap to next service                                        
[EPGRefresh] Done refreshing EPG                                          
<578997.568603> [eEPGCache] store epg to realpath '/media/ET10000/epg.dat'    
<578997.737527> PC: 76cc3870                                                  
<578997.737642>     00000000 00000001 74dafc99 00000001                       
<578997.737707>     74dafc99 00000000 00000364 00000003                   
<578997.737768>     82b54a76 a7622e5d 00000000 f292d4ed                       
<578997.738214>     1de1e7df 2d855de9 09505640 2ce4f084                       
<578997.738308>     00000367 02532be8 000003fc 00000000                       
<578997.738368>     00000000 00000000 000120e3 0081d924                   
<578997.738434>     00000001 76cc36f0 769fc084 00000000                   
<578997.738494>     76db2e30 7f87f428 73815130 76cc30a8                   
<578997.738611> -------FATAL SIGNAL                                       
]]>

All crash log files have the same time of day:

root@et10000:/media/hdd# ls -la
drwxr-xr-x    4 root     root          4096 Jun 30 04:53 .
drwxrwxrwt    6 root     root           120 Jan  1  1970 ..
-rw-r--r--    1 root     root         17415 Jun 21 04:53 enigma2_crash_1434855212.log
-rw-r--r--    1 root     root         17415 Jun 23 04:53 enigma2_crash_1435028012.log
-rw-r--r--    1 root     root         17415 Jun 25 04:53 enigma2_crash_1435200812.log
-rw-r--r--    1 root     root         17415 Jun 27 04:53 enigma2_crash_1435373612.log
-rw-r--r--    1 root     root         17415 Jun 28 04:53 enigma2_crash_1435460012.log
-rw-r--r--    1 root     root         17415 Jun 29 04:53 enigma2_crash_1435546412.log
-rw-r--r--    1 root     root         17415 Jun 30 04:53 enigma2_crash_1435632812.log
drwx------    2 root     root         16384 Oct  9  2009 lost+found
drwxr-xr-x    2 root     root         98304 Jun 30 04:34 movie

As you can see it does not happen every night. There are nights when EPG refresh is running without a crash.

After the enigma2 restart It looks like that EPG refresh is running again successfully this time. The time stamp of epg.dat shows 05:47:

root@et10000:/media/ET10000# ls -la
drwxrwxrwx    6 root     root          4096 Jun 30 04:53 .
drwxrwxrwt    6 root     root           120 Jan  1  1970 ..
drwxrwxrwx    2 root     root          4096 Jun 30 07:30 backup
-rwxrwxrwx    1 root     root             0 Dec  4  2014 backupstick
-rwxrwxrwx    1 root     root      14395328 Jun 30 05:47 epg.dat
-rwxrwxrwx    1 root     root      14395328 Jun 30 05:47 epg.dat.backup
drwxrwxrwx    2 root     root          4096 Jun 18 20:59 et10000
drwxrwxrwx   11 root     root          4096 Jun 18 20:59 fullbackup_et10000
drwxrwxrwx    2 root     root          4096 Jun 18 12:09 ipk
-rwxrwxrwx    1 root     root         21504 Jan  1  1980 skydb.db

Could please someone have a look into this?

[socks]

There is one more thing I discovered.

I had a look into the list of completed timers.

On June 22nd, 24th and 26th when no crash occured I had scheduled recordings around 4am.

This means there are only crashes when box is in standby and no recording is running.

EPG refresh is configured to run even when box is in use.

[littlesat]

Looks like something invalid in the epg.dat. The box does crash on it.

Edited by littlesat, 1 July 2015 - 06:38.

[Dimitrij]

Previously, no such problems.
Corrupted files simply removes.

[ims]

Did you add some new service for refresh ? F.eg 19.2E - Sky Select ?

[socks]

Did you add some new service for refresh ? F.eg 19.2E - Sky Select ?

 

All my user bouquets are in the refresh list.

Sky Select is included in one of the bouquets.

I can not tell exactly when the crashes began to happen but I have not added any services recently.

Now I removed Sky Select .

Unfortunately a manual EPG refresh always succeeds so I have to wait until tomorrow to see what happens.

[ims]

remove this Sky Select from refresh ( create own bouquet for it and do not put it then into refresh). This making problem... subservices...

 

And may be - better would be: init 4, remove epg.dat and then init 3

Edited by ims, 1 July 2015 - 10:44.

[socks]

remove this Sky Select from refresh ( create own bouquet for it and do not put it then into refresh). This making problem... subservices...

 

And may be - better would be: init 4, remove epg.dat and then init 3

 

AFAIK EPG data is the same for all services on one transponder. Because of that most of the service on my refresh list are skipped during EPG refresh.

So in my opinion it will make no difference only removing Sky Select.

Nevertheless I will give it a try.

 

I already tried init 4 / rm epg.dat / init 3. It did not work. Same crash next night.

Edited by socks, 1 July 2015 - 11:15.

[ims]

yes, but subservices are not standard ...  I am 100% sure that it is Sky Select !

[socks]

Same crash this night with Sky Select removed:

<751679.978284> [eDVBSectionReader] DMX_SET_FILTER pid=57                  
<751686.979053> [eEPGCache] abort non avail schedule other reading         
<751686.979688> [eEPGCache] abort non avail netmed schedule reading        
<751686.980384> [eEPGCache] abort non avail netmed schedule other reading  
<751686.980743> [eEPGCache] abort non avail FreeSat schedule_other reading 
<751686.981239> [eEPGCache] abort non avail viasat reading                 
<751694.984461> [eEPGCache] nownext finished(1435805510)                   
[EPGRefresh] Maybe zap to next service                                     
[EPGRefresh] Done refreshing EPG                                           
<751737.583229> [eEPGCache] store epg to realpath '/media/ET10000/epg.dat' 
<751738.006708> [eEPGCache] 159805 events written to /media/ET10000/epg.dat
<751738.431323> PC: 764f9848                           
<751738.431433>     00000000 00000001 761c100b 00000001                    
<751738.431499>     761c100b 00001000 00000004 761c100f
<751738.431566>     00000000 819fe5c0 00000001 00000118
<751738.431630>     fffffffa 00000000 00000001 08000000                                             
<751738.431691>     00000004 025e5b50 00000004 00001000
<751738.431750>     00000000 00001000 0002703d 0081db64
<751738.431809>     00000000 764f96f0 76232084 00000000   
<751738.431868>     765e8e30 7fb1b380 00001000 764f90a8
<751738.431994> -------FATAL SIGNAL
]]>                               
                </enigma2crashlog>
        </crashlogs>
</openpli>

[ims]

Are you sure, that you remove epg.dat before it ? Has not this service in some next bouquet ? Is not created autotimer from this service ?

[socks]

Yes, I'm sure.

I removed epg.dat and epg.dat.backup (init 4, rm, init 3). There was no such file when EPG refresh started.

Sky Select is not present in any of the user bouquets.

I have not switched to this channel. Nevertheless EPG data for it is stored because there are other channels using the same transponder and sharing same EPG data.

There is an option to start the autotimer search when EPG refresh has finished. This option is disabled.

There are no autotimers using this channel.

I will now try to remove user bouquets from the refresh list one by one to see which one is causing the problem.

[ims]

I had this problem, when I add Sky Select (due adaptation fullhd skin for it) into bouquet, which was refreshed - from this time started gsod ... 2 type - one crash when was refreshed EPG and data from epgcache was saved. Second crash type was periodicaly by autotimer, when looked once a day in epgcache.

After removing Sky Select it is all ok again.

 

May be, exist next program with subservices in your bouquet ? As I remember, some years ago there exist 2 channels "Premiere". One for movie, one for sport.

[Dimitrij]

normal save epg

<  2529.043183> [eEPGCache] 6781 events written to /media/usb/epg.dat
<  2530.651187> [eDVBLocalTimeHandler] set RTC to previous valid time
<  2530.651889> [eDVBFrontend] close frontend 0

[eDVBLocalTimeHandler] set RTC to previous valid time

this

<751738.431323> PC: 764f9848                           
<751738.431433>     00000000 00000001 761c100b 00000001                    
<751738.431499>     761c100b 00001000 00000004 761c100f
<751738.431566>     00000000 819fe5c0 00000001 00000118
<751738.431630>     fffffffa 00000000 00000001 08000000                                             
<751738.431691>     00000004 025e5b50 00000004 00001000
<751738.431750>     00000000 00001000 0002703d 0081db64
<751738.431809>     00000000 764f96f0 76232084 00000000   
<751738.431868>     765e8e30 7fb1b380 00001000 764f90a8
<751738.431994> -------FATAL SIGNAL

service  is none,not valid time

[Dimitrij]

Yes, I'm sure.

I removed epg.dat and epg.dat.backup (init 4, rm, init 3). There was no such file when EPG refresh started.

Sky Select is not present in any of the user bouquets.

I have not switched to this channel. Nevertheless EPG data for it is stored because there are other channels using the same transponder and sharing same EPG data.

There is an option to start the autotimer search when EPG refresh has finished. This option is disabled.

There are no autotimers using this channel.

I will now try to remove user bouquets from the refresh list one by one to see which one is causing the problem.

What's next?
There are the results?