Big Release... 1.8.0.0 is here.

[Space]

I'm pretty sure it indexes the WMC DB too, as people in the past have had problem with WMC not scheduling programs to run, and it turned out the main Windows indexing was either not running or was corrupted.

Plus you can look on the page I mentioned above ('Indexing Options") and see that one of the entries indicates to index "Windows Media Center".

[IT Troll]

Whilst it is listed as an indexed location, the Media Center indexing is performed as two separate tasks (as Gary mentioned) with separate schedules to that of the file system indexing. The PVR schedule indexing is normally very quick and doesn’t in itself trigger indexing of other things. Could it be that other background tasks just happen to be occurring at the same time?

It would be interesting to run the pvrschedule manually from the command line at a time when the drive/system is otherwise idle.

A mechanical drive is a real bottleneck for a system drive and database though. A small SSD for Windows, WMC DB, indexes, swap files, etc. would make a massive difference. You can still use the mechanical drive for recordings.

[Space]

Since my system is up 24/7, it is highly unlikely that any other indexing would be taking place after a guide update other than any files that were touched by the update itself. As far as I know, indexing of files is done soon after they are modified, and not on any kind of regular schedule.

[garyan2]

I think we might be getting off track here.

This is what I think happens and what I may need to do.

The fact is there are 2 indexing tasks.

The first for the guide, indexing all the listings, descriptions, people, keywords, etc.... This takes a while depending on database size.

The second is strictly for recording requests. This is done automatically at the end of the first indexer. It can be done independently. I believe it may be done after anytime you add/remove/alter any recording requests. This is very quick unless you have a database problem or have hundreds of scheduled recordings that WMC has to prioritize for tuner conflicts.

Based on all that, if the PVR indexing needs the guide indexing to complete for it to be accurate, then that means I would have to monitor the indexing. What do I do if someone's Windows Search feature is broke and WMC is not getting indexed? What do I do if it takes hours to complete indexing?

What I think I may do is evaluate the recording requests after importing the mxf file and performing a quick PVR indexer; before the guide indexing starts. I will need to monitor the status of the PVR indexer prior to evaluating for tuner conflicts but should be quick. I might as well move the recorder storage evaluation to the same point in the process.

[Space]

If I had to guess, in order to get proper PVR indexing, you would have to wait for the guide indexing to complete. I would think that the PVR indexing is dependent on having a properly indexed guide and is the reason why PVR indexing happens automatically after guide indexing.

If this is the case, then running PVR indexing immediately after loading the MXF, and before running guide indexing would have no effect.

Again, this is total speculation based on nothing but what seems logical to me, and of course may be completely wrong. There may actually be no need to run guide indexing before WMC can recognize new conflicts caused by the recent MXF import.

[garyan2]

I'll have to run a test. Certainly it is a possibility that the database reindexing has to be complete, but I don't think it is very likely.

When new guide listings are imported, you can see them right away in the guide. Everything is in the database and linked correctly. That should be enough for the PVR indexing to be complete, successful, and accurate. What you won't get until after the database reindexing is complete is any of the newly added programs showing up in any search results within WMC.

Again though, need to run a test to verify one way or the other.

[IT Troll]

I believe file indexing begins when the system appears idle. I turned off indexing of many folders as it is not required for a HTPC. Certainly don’t bother indexing recordings or other media files.

I guess a fringe case might be someone who has setup a keyword based recording. That might need a full guide index and might generate a new scheduling conflict. I think that feature is sometimes used by folks for sporting events.

I find having a quad tuner is the best way to avoid conflicts. It would be even better if Media Center could record the whole mux as other PVR systems can.

[Space]

...
I guess a fringe case might be someone who has setup a keyword based recording. That might need a full guide index and might generate a new scheduling conflict. I think that feature is sometimes used by folks for sporting events.
...

Good thinking there, you are probably right about this, even if detecting Series and one-off recording conflicts don't require the guide to be indexed, this type of recording is much more likely to require it.

[garyan2]

Yes, good thinking and I will say my test last night failed. It wouldn't pick up a "new" conflict until after the database indexing and PVR indexing were both complete. My computer was acting a little funny yesterday and my database indexing took well over an hour to complete for some reason (I gave up and went to bed). I'll try again soon.

[garyan2]

...and yes, certainly with keyword searches the database reindexing needs to be completed prior to checking for conflicts.

[Space]

One suggestion I have is after you finish the guide import and kick off the indexing, and before the client exits, that you also start a separate watchdog indexing monitor process that monitors the indexing progress. In this way it does not block any EPG123 client activity initiated by the user (by not allowing more than one to run at the same time) but still allows for monitoring of the indexing process and reporting of any conflicts.

This watchdog would monitor the indexing, and when the indexing is completed, check for any conflicts and send a notification if any are found. This watchdog should also have a timer that if the indexing has not completed within a certain period of time, a notification is sent that there may be an issue with indexing.

I think this timer should be configurable, as different systems have different reasonable times for completing indexing. Also, if another instance of the watchdog is started, it should simply update the existing timer on the existing watchdog and exit (or replace that instance with itself, which would also reset the timer). This would prevent the issue of having multiple watchdogs running at once if the user initiates guide updates one after the other and the indexing takes a while to complete.

I'm not sure if there is a way to monitor the WMC specific indexing as opposed to just general system indexing, but if there is, you should probably only monitor the WMC specific indexing.

[IT Troll]

Or more simply, how about two schedule tasks. One for the update/import that we know and love, then one for "health checks" like disk space and conflicts. People can then set the times that would best fit with their system's operation and recording routine, or even choose not to schedule the health check if they are not bothered about email alerts.

[garyan2]

Or more simply, how about two schedule tasks. One for the update/import that we know and love, then one for "health checks" like disk space and conflicts. People can then set the times that would best fit with their system's operation and recording routine, or even choose not to schedule the health check if they are not bothered about email alerts.

I was thinking something along those lines, but more like setting up another task that would trigger upon completion of the database indexing task. It would require turning on task history though. Still brainstorming....

[garyan2]

Just released 1.8.0.12 that performs the storage and tuner conflict checks after the database reindexing has completed. There will be log entries, email notifications if setup, and will change the tray notification and status logo if necessary.

No new tasks or anything. Just running another session of the client in the background to kickoff and monitor the indexing.

[Space]

Nice!

One suggestion I have for something I noticed is that you add a line to the log file to indicate that the "WMC recorder storage/tuner conflict checks" have completed, similar to what you do for the automated update and client execution:

Example existing log entry:

Code: Select all

[7/16/2023 6:53:46 PM] [ INFO] Completed EPG123 update execution. SUCCESS.

Possible log entry:

Code: Select all

[7/16/2023 6:53:46 PM] [ INFO] Completed WMC recorder storage/tuner conflict checks. SUCCESS.

I was monitoring the log, waiting for it to finish, not knowing that it had already finished...

I also noticed there was no log entry indicating it was checking for conflicts, or the result, I guess that is fine as long as it is actually doing those checks.

I also noticed that the indexing finished and then the client noticed about 2 or 3 minutes later, how often is it polling for the indexing to be completed?

[garyan2]

There should be an entry in the log like the following:

Code: Select all

[7/16/2023 1:04:02 PM] =====================================================================================
[7/16/2023 1:04:02 PM] EPG123CLIENT: Beginning WMC recorder storage/tuner conflict checks. version 1.8.0.12
[7/16/2023 1:04:02 PM] =====================================================================================
[7/16/2023 1:04:02 PM] *** Windows 11 Home, 64-bit [Version: 10.0.22621.1992] (22H2) ***
[7/16/2023 1:04:02 PM] *** .NET Framework 4.8.09032 is installed. ***
[7/16/2023 1:04:02 PM] *** Windows Media Center [Version: 6.3.9600.16384] is installed. ***
[7/16/2023 1:04:02 PM] [ INFO] WMC database indexing took 00:06:42.8236044. Exit: 0x00000000
[7/16/2023 1:04:02 PM] *** Recorder storage drive C:\ has 223.561 GB available. (7.192 GB of 230.753 GB used) ***

There is no polling. The second instance of epg123Client starts the indexing using ehPrivjob.exe and waits for it to complete.

[Space]

Yes, I see those log entries, just thought there should be a "completed" log entry for the storage/conflict checks like the update and import tasks have.

[garyan2]

Understood. Kinda curious on why you say there was 2-3 minutes after indexing was done that the checks were started though.

[garyan2]

So here is what next release will look like.

Code: Select all

[7/16/2023 1:04:02 PM] =====================================================================================
[7/16/2023 1:04:02 PM] EPG123CLIENT: Beginning WMC recorder storage/tuner conflict checks. version 1.8.0.12
[7/16/2023 1:04:02 PM] =====================================================================================
[7/16/2023 1:04:02 PM] *** Windows 11 Home, 64-bit [Version: 10.0.22621.1992] (22H2) ***
[7/16/2023 1:04:02 PM] *** .NET Framework 4.8.09032 is installed. ***
[7/16/2023 1:04:02 PM] *** Windows Media Center [Version: 6.3.9600.16384] is installed. ***
[7/16/2023 1:04:02 PM] [ INFO] WMC database indexing took 00:06:42.8236044. Exit: 0x00000000
[7/16/2023 1:04:02 PM] [ INFO] Recorder storage drive C:\ has 223.561 GB available. (7.192 GB of 230.753 GB used)
[7/16/2023 1:04:02 PM] [ INFO] No tuner conflicts detected within next 3 days.
[7/16/2023 1:04:02 PM] [ INFO] Completed WMC recorder storage/tuner conflict checks.

[Space]

Understood. Kinda curious on why you say there was 2-3 minutes after indexing was done that the checks were started though.

I was monitoring the indexing on the Indexing Options screen and when it changed to "Indexing complete" it read 7:19 on the system clock (I didn't check the seconds). In the log it logged that "WMC database indexing took 00:10:32.64125778" at 7:21:50. So it was about 2-3 minutes after the Indexing Options screen indicated it was completed.