49 replies to this topic

[allstarnz]

got pointed in this direction by NFSNovice...

I have a DM7020 using Jade and SQLite and have exactly the same thing. Seems to have crept in this latest build perhaps? I was running the build prior to Jade final, did the updates a few weeks ago and bam.

As suggested in this thread, I have changed the timezone, and set a swapfile of 8MB on my USB stick. Timezone change seems to have made a difference, faster but still not as good as it was not that long ago.

My EPG is 27.1MB

[WanWizard]

Another tip that I have a bit of success with at the moment:

Analysis of the database file has shown that db_epg (the import tool) has a negative impact on the way the records are stored in the database, leaving large gaps amounting up to 70% of the file. SQLite has a command called VACUUM, which reorders, compacts and reindexes the database. We have this command scheduled in Enigma every night at 04:00. I see a significant improvement in the stability and performance of the box if I schedule the radiotimes import so that it is finished before 04:00, when Enigma reorganizes the database.

Someone willing to try this as well, and see if the results can be reproduced?

[nfnovice]

could even put the vaccuum (reorg and pack) as the last line of the epg import script - load it up and then clean it up.
what routine looks after throwing out previous days entries - that would leave gaps in the database anyway

[WanWizard]

True. We've looked at this, but a VACUUM takes quite a lot of time (on my box a little over 3 minutes last night), and while it's running the box is unresponsive. It is therefore not a good idea to have enigma do it at regular intervals.

It might be an option to have db_epg do this, but I'd like some confirmation first that this really helps.

I just checked the logs of my debug version, and I only see a DELETE query when Enigma starts. So it looks like there is no regular cleanup of old records. This also might have a negative effect if you never restart Enigma. We'll have to look into this as well.

[nfnovice]

I would think that most people run the scripts to load epg in the wee hours of the morning - that would probably also be a decent time to run any maintenance on the epg.
I note your comments on the the delete query only running when enigma starts - interesting.
What about the option to limit epg store size in EPG settings - what effect does that have - logic would suggest it must enable some deleting ?

[WanWizard]

The limit epg store size is a setting for the standard memory EPG. When you restart enigma, this EPG is dumped to flash (default), and when it is to big, you get storage problems or enigma refuses to start. The americans have a lot of problems with that, Dish/BEV send lots of EPG data. It isn't used for SQLite.

With regards to the delete query, other team members have reported that their box runs delete queries on a regular basis. We're trying to find out why mine doesn't.

I have now scheduled the epg scripts at 03:30, before the maintenance, and that seems to work pretty well. The box is still responsive, and hardly no hickups noticable when browsing though the service selector. Still early days to call this conclusive though...

[allstarnz]

my EPG was playing silly buggers, so I killed the database, and started again. Goes much faster now.

I tried vacuum in two different times, both times it said the database was malformed.

[Sicilian]

I'm willing to try vacuum if the details can be posted please.

[WanWizard]

Enigma does a VACUUM every night at 04:00, so there's nothing you should do manually.

What helps in my case (YMMV) is to run the import script before 04:00, instead of after it, so that the database can be compacted after the import.

[doggone1]

Hi all just found this thread after 3 months of freezing and lock ups even remote control was unprdictable even after updating to latest version of jade set to system time and alls working great no problems at all thanks for posting this