Sudden spinning wheel when show ends

[jhkoenig]

This one really has me scratching my head. When a show ends, the listing of recorded shows (or imported movies) appears in the UI as normal. Pressing any key on the remote (like an arrow or delete key) triggers a spinning wheel that takes 5-20 seconds to disappear. If I press Enter during the spin, intending to delete the show I just watched, the dialog box that pops up is as though I was in the left pane, so it presents "Browse Recordings" as the first entry instead of "Watch Now". If I "Close" out, I have the select bar positioned correctly and another Enter press shows the proper dialog box.

Everything else works. I have a huge Java heap, so it isn't garbage collection, and all the disks have 100% SMART health reports. The only error I see in the Sage logs is a consistent failure to look up "2021 Olympics" as a show name.

Any ideas? Thanks!

[stuckless]

Do you have a multi-disk system? This is an exact problem that I had when I first setup an unRAID system. Basically, videos played fine but when the video ended, and I went back to the sagetv UI, and then did anything, I'd get a spinning circle. Turns out that my drives were spinning down, and when the UI needed to render something new it had the load the artifacts from the disk and the disks would need to spin up, taking 5-20 seconds.

After the spinning circle goes away, are you fine after that? (I suspect so).

[jhkoenig]

Originally Posted by stuckless

Do you have a multi-disk system? This is an exact problem that I had when I first setup an unRAID system. Basically, videos played fine but when the video ended, and I went back to the sagetv UI, and then did anything, I'd get a spinning circle. Turns out that my drives were spinning down, and when the UI needed to render something new it had the load the artifacts from the disk and the disks would need to spin up, taking 5-20 seconds.

After the spinning circle goes away, are you fine after that? (I suspect so).

Thank you for your reply! My disks are not spinning down. In fact, this happens even when I am recording a show to each of my 6 drives, ensuring that the drives are spinning.

Once the spinning circle stops, I am good to go until the next "watch, stop" cycle.

Is there such a thing as too much Java heap? I bumped it way up when I was fighting Java GC. It is sitting at 6GB right now.

[stuckless]

Originally Posted by jhkoenig

Thank you for your reply! My disks are not spinning down. In fact, this happens even when I am recording a show to each of my 6 drives, ensuring that the drives are spinning.

Once the spinning circle stops, I am good to go until the next "watch, stop" cycle.

Is there such a thing as too much Java heap? I bumped it way up when I was fighting Java GC. It is sitting at 6GB right now.

Basically you'll see spinning circles for...

• slow disk IO (spin downs, other processes thrashing the disks, etc)
• bad network (as was the case a few days ago for someone here)
• java gc

If you can rule out disk IO, then I suspect the network is OK (you'd usually see other things like sluggish UI on clicks), you are left with GC. Yeah it can cause this. I think think about how Java works (although different from release to release) it will just eat memory until it runs out and then it will GC. So, if you give it 6G and it doesn't claim memory until it reached that, then it will then have to process 6G worth of memory to clean it up. With Java 8+ there are other kinds of GC options you can provide to Java, but I have no idea how to set them on windows.

[jhkoenig]

So this spinning wheel issue seems to be somehow related to rendering a dual-pane list of recorded shows or imported videos. I have seen the spinning wheel happen by pressing channel/Page up after waiting for a spinning wheel to expire! Makes the Java GC theory hard to support. Is there some magic number of shows/episodes that causes Sage to pause and count on its fingers? With the Olympics going on and 6 tuners, I have a lot (about 80) "episodes" of the Olympics freshly recorded and hundreds of other shows, a few with over a hundred episodes.

This is driving me nuts! No idea is too crazy to explore.

[jhkoenig]

Oh, another clue: this issue to not affect a Win10 client. No spinning wheel at all!

[stuckless]

SageTV will log when it runs into spinning circles (I think it logs something like UI HANG in the log). You can check to see what it is doing just before that.

Have you tried with just the default STV?

Have you checked your network (see the previous discussion on slow UI)

Does it only happen on your olympics videos? maybe you have a bad video in the list.

I would expect a full client to be fine in most cases.

[jhkoenig]

Originally Posted by stuckless

SageTV will log when it runs into spinning circles (I think it logs something like UI HANG in the log). You can check to see what it is doing just before that.

Have you tried with just the default STV?

Have you checked your network (see the previous discussion on slow UI)

Does it only happen on your olympics videos? maybe you have a bad video in the list.

I would expect a full client to be fine in most cases.

Thanks for the ideas!

The log as 1,200 HANG lines. Example:
Sat 2/19 14:04:34.970 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 751 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:35.720 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 1501 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:36.471 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 2252 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:37.222 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 3003 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:37.972 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 3753 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:38.722 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 4503 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:39.472 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 5253 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:40.223 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 6004 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:40.973 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 6754 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:41.724 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 7505 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:42.475 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 8256 UILocker=Thread[ReProcessHook,5,main]
Sat 2/19 14:04:43.225 [AWTThreadWatcher-0023a5000333@3d879109] EventThread-0023a5000333 Hang Detected - hang time = 9006 UILocker=Thread[ReProcessHook,5,main]


But I don't know what this indicates. The network is hardwired and solid (my career has been in IT) so I doubt a network problem. I also learned that deleting a show is another easy way to trigger the wheel, so a playback sequence is not even required.

There were two "miracle moments" yesterday when the issue did not appear, but 10-20 minutes later it was back. Getting close to rebuilding the Sage instance from the ground up. Maybe some plugin left entrails behind after an uninstall?

[stuckless]

What is happening just before those HANG lines, since that is likely the last think sagetv was doing before it hung.

Thist still looks like an HD IO issue to me. either disks are spun down, OR bad media file that sagetv is having a hard time with. You have 6 drives, and it only takes one of them to spin down and sagetv to try to access some file on it (fanart picture, or properties, etc) and you are blocked until it comes back. This was in fact my problem. My media was spread about 4 drives and depending what list of media I was trying to load after watching a show, it would spin, while it was accessing some piece of fanart for some show it was trying to show in the list. I don't know much about windows, but, if there's an option to say "never spin down" for your drives, then enable that for a few days to rule out the drives spinning down.

I agree that it is less like an network issue, and given your experience, I think you are rule it out.

6GB JVM is massive for 3 extenders. I would put it around 3-4 max.

I did ask if you were using just the default STV (or at least tried with default stv).

If this is JVM or HD IO, then a full rebuild is likely not going to do much.

Maybe try a different java version 64bit Java 11.

[jhkoenig]

Thanks again, Stuckless, for your latest reply! This forum would be poorer for your absense.

I don't know how to interpret the "hang detected" messages. Is there a way to back trace these messages to the action which is hanging? Further, is there a "dead man's switch" capability kill processes that run beyond some set hang time?

I like your thinking re fan art or properties triggering the hangs. I can try to move the fan art to a different drive. Does Sage buffer incoming shows through the "home" drive before moving to the targeted drive, or is it a direct shot? My home/boot drive is an SSD and I'm wondering about deterioration on that which is invisible to SMART monitoring.

To make things more fun, this problem has become intermittent since the Olympics are over, so making changes and testing are now more challenging.

Thanks again, Stuckless, I'll chase your ideas and report back!