[Usability] Restart on resume (1 Viewer)

[Requiem]

Hi,

I'd like to re-initiate a discussion about a feature/workaround which has been discarded with the release of MP1.4: The option to automatically restart MediaPortal on Resume from Standby.

The functionality always has been a workaround for existing bugs, and of course, it's still a valid argument that it would be better to eleminate any Bugs in MP that prevent a clean startup after resume than to use a workaround. Yes, I understand that.

Well, my reality is as follows: I have a multiseat-setup (one server, one main HTPC, AmpDroid and two notebooks...). I always try to use the latest mediaportal version (1.8pre right now) for more than 2 years now.
When everything goes wrong and I have a f*cked up configuration or irreproducible errors, I'll just go back to a clean Windows7-Image with all updates of my HTPC and reinstall MP from scratch with all its plugins.
Right now, some very essential plugins lack official releases, like Fanart Handler.
So, I install MP, I install my favorite skin with its plugins (StreamedMP), I configure everything, I start with newly built databases for fanart, series, music etc.
At the moment, I need to find time again to reinstall from scratch once again, as with 1.7.0/1.7.1hotfix/1.8pre I have issues with resume again - this time EVERY time when I resume, I have to kill MP as it's frozen or just hit the HTPCs reset button.

I keep reading several posts in several threads here, where the technically advanced users are sharing their problems with MP and resume from standby. Just search the forums yourselves.

As I said, I fully understand that the goal is to eleminate bugs and not to implement workarounds.

But please, reimplement this functionality or implement it as an addon. It would save me (and I am sure, lots of others) hours of frustration. There still seem to be a lot of circumstances not obvious to the standard end-user who just wants to use his HTPC which lead to a state of not responding/crashing MP after resume from standby.

So, a pledge to the devs: Please rethink your arguments for and against the restart on resume functionality. I still think that MP would be better with this workaround feature.

 

[Requiem]

Maybe there's something wrong with my system, but I cannot figure it out. I have reset my Windows PC to a fresh installation start and made a client-only-installation of 1.8pre with StreamedMP, configured everything and everything works - except a clean resume from standby.
Day 1: MP hangs after resuming from standby (black screen, no reaction)
Day 2: Corrupted database after resume from standby, needed to copy it back from my incremental backup after clean installation
Day 3: MP manages it to come back properly after standby (I didn't change anything)
Day 4: MP hangs again after resume from standby.

I'd write a plugin to restart after resume if I knew how...

 

[Luca Brasi]

Back in the day when I had problems with this as well I used Eventghost to close MP before going to sleep and start it after resume. Be aware that Eventghost has it's flaws as well and that you will have to dig into this one also. When you have it running as it should it is a mighty tool though.

Edit: I just realize this is in the development section so my hint might not fit in here. If so, please just ignore it.

 

[mm1352000]

Hello Requiem

I remember several discussions on this subject in which you've strongly advocated for us to bring back this function. Your points and logic have always been well thought out, and I think you've also fully understood the main response from us about avoiding workarounds.

Here's the thing that is troubling me: I've gone through your post history and I can't find a single set of logs from you to enable us to try to solve the standby/resume problems you're experiencing. Maybe I missed a post somewhere, I don't know... but how about giving us [another] chance to look at your problem with fresh eyes? While I understand and even agree that having the function back could be useful for more people that just you, I also think that solving your problem could solve the problem for more people than just you... and solving the problem would be a better solution because no configuration would be required.

Regards,
mm

 

[Requiem]

@mm1352000 Thanks for your interest in this topic. Well, yes, I think both of us understand the other one. I am gladly willing to help to participate to solve any problems where I can. I think I have posted logs about this before, but don't nail me down on this. Anyhow, I'm not at home right now, but I will collect a set of logs in the next days and will gladly provide them.

One thing beside hangs is a problem with database corruption, I am not sure why this is happening; I've had corrupted databases where MePo wont start anymore a bunch of times and was able to solve this problem by restoring the last backup of the database folder (I do automatic incremental daily backups).

But would it be difficult to take the old code and "export" it to an addon to be able to restart on resume until there are no more issues?

@Luca Brasi: Thank you, I'll loook into this eventually, but seems to be quite complicated. I dont need EventGhost for other purposes right now.

 

[HomeY]

One thing beside hangs is a problem with database corruption, I am not sure why this is happening;

Do you happen to use the Central database option and modified your MediaPortalDirs.xml?

and was able to solve this problem by restoring the last backup of the database folder

This would mean it's plugin related, since you're talking about the SQLite databases, and not the TV database (which uses MySQL by default). Correct?

 

[Requiem]

@HomeY No, I don't use a central database modification as I sometimes have more than one client accessing. I use MPSync's current version to synchronize from my HTPC to the server (TV-Server and media storage) to have an up-to-date MPExtended database fpr use withe WebMediaPortal and aMPdroid.

As I have a multiseat setup, it's only the clients databse folder which has seen corruption so far (for the server, everything goes well with resume from standby using LightsOut-Software, the HTPC will wake the server after resume through the lightsout-client software, MePo client is configured to wait a few seconds after resume to give the server time to come online). I have seen corruption with plugindatabases as Moving Pictures or MP-TV-Series, but as well with the music database.

 

[HomeY]

I'm interested in the following test:
Disable the MPSync plugin before going to bed tonight, then start MediaPortal with the WatchDog option #2, let (or put) the system (to) sleep and see how your databases are doing tomorrow morning (or after resume).

But let's wait for logs first before i keep guessing

 

[Requiem]

I'll do so. But though database corruptions happens from time to time, this is not my daily problem with resuming, just to be sure.

 

[HomeY]

But though database corruptions happens from time to time, this is not my daily problem with resuming, just to be sure.

Yeps, got that. But since you stated in your OP that you can reproduce it every day, the logs should tell us more.
For us it's important to know if it's MP Core or Plugin related, because that's about the first step to narrow the issue down to it's source.