Long delay opening TV Collection

[pgriffith 65]

276 for me 'seems' to have lessened the occurrences.

 

I went in and out of 30 tv shows, it only happened twice this time around.

 

Tried number 31 while typing this, went in immediately, but there was no content, I left it for awhile and hit the back button once, nothing happened, left it a bit more than got the Program Not Responding dialog pop up. Left it alone some more and the content appeared behind the error box. Clicking cancel allowed me to continue on. 32 is currently doing the same thing .... sigh

 

Edit.. Scratch the above lessened occurrences

 

Nearly every 2nd show or so from 30 onwards has had the delay

Edited June 8, 2015 by pgriffith

[ebr 14913]

277 going up with a bunch of logging in it.

 

Pgriffith there is a chance this version will fix the problem for you but, if it does, it isn't a real fix for most.

[pgriffith 65]

I was just going to bed too  (2am here)

 

Just one quick test  

[MrWebsmith 598]

277 FIXES IT...welll so far anyway.. im not able to reproduce anymore...

 

logs coming anyway just for analysis of the new msgs..

 

 

mbc 277 log

 

https://www.dropbox.com/s/t290eqh9vpxqcp1/MBClassic-86201599665c612a2b4efca92db85267bfda2d.log?dl=0

 

emby server log updated:

 

https://www.dropbox.com/s/v7g0kn2tlhwmtoc/server-63569352156.txt?dl=0

 

 

i was all over the place in this log... tryign to get it to fail/delay... so a number of shows.. older/newer/3rd rock/24... you name it i was really pushing on it...

 

more testing will take place tonight when we go to watch shows.. GoT! Falling Skies!

 

 

thanks ebr.. as always..

Edited June 8, 2015 by MrWebsmith

[ebr 14913]

Web - are you running without the special TV view as well?

 

If so, can you update to 278 and turn that view back on and then try and reproduce?

[pgriffith 65]

 

 

55 Shows entered and SUPER fast, with all content showing... don't think it's ever been this responsive.

 

Care to say what you did in laymans terms ?

 

Thank you so much.

[MrWebsmith 598]

ebr

 

yup i have that disabled/unchecked.. AND im a "use legacy folder browsing" enabled user.. 

 

choc views>"use custom view for tv series"...unchecked..

 

 

i will up to 278 and retest to compare

[pgriffith 65]

with 278 and the use custom view for tv series enabled, I am getting delays on some shows again. Happened on 5 out of 30 shows, some had the long delay on getting the episode content listed.

 

 AND I'm a "use legacy folder browsing" enabled user as well

 

I'll try again with custom views off to make sure it wasn't something that was particular to 278

[pgriffith 65]

278 with custom views OFF seems to be fine. Entered and exited 40 shows not a single glitch.

[MrWebsmith 598]

agree with pgriffith..

 

custom view ON, delays are back on 278... custom view OFF.. working as well as 277 was (immediate/no issues)

[pgriffith 65]

I'm going to bed now, will pick this thread back up in the morning    (so excited, I have 3 other users with this issue as well)

 

Cheers guys

[ebr 14913]

Yeah, I thought this would solve it if you have the custom view turned off but it isn't really solved - just doesn't apply to that situation anymore.

 

Web, can you please reproduce the problem with 278 and then post the MBC log.  Thx.

[MrWebsmith 598]

ok in order to do that i need to turn ON the custom view.. just to confirm this is what you want to see

[MrWebsmith 598]

https://www.dropbox.com/s/0tyjucaz1mges3z/MBClassic-8620152de112bcc9de4d618ddc10094c30bd04.log?dl=0

 

278 log same series (3rd rock...) 3 entries.. 1 fast 2 delayed..3 delayed

 

let me know if emby server log should also be updated/reposted

 

emby server log updated:

 

https://www.dropbox.com/s/v7g0kn2tlhwmtoc/server-63569352156.txt?dl=0

Edited June 8, 2015 by MrWebsmith

[ebr 14913]

Thanks but I still can't be sure exactly about the client/server interaction because your server is reporting that it received the request  a minute before the client is sending it...  Are you sure about those system times...?

2015-06-08 14:23:37.0323 Debug - HttpServer: HTTP GET http://192.168.1.15:8096/mediabrowser/Users/f23a949b-5da2-dd2c-b526-bf4fa8738d39/Items?ParentId=1c03d9463c1bee13f6676676918ce5b8&sortBy=&CollapseBoxSetItems=True&fields=Overview,Path,ParentId,DisplayPreferencesId,DateCreated,IndexOptions,DateLastMediaAdded,Metascore,MediaStreams,SortName,Taglines,MediaSources,SpecialEpisodeNumbers&ImageTypes=&recursive=False&genres=&Ids=&studioIds=&ExcludeItemTypes=&IncludeItemTypes=&ExcludeLocationTypes=&personIds=&personTypes=&years=&format=Json
	Ip: 192.168.1.23:52762. Headers: Authorization=MediaBrowser UserId="f23a949b-5da2-dd2c-b526-bf4fa8738d39", Client="MB-Classic", Version="3.0.278.0", DeviceId="15bb04c0d02e4832acbbfc2cdd26f31a", Device="HTPC-NUC/Matt",X-MediaBrowser-Token=87da5d65ebc3439fac788c69775f238e,Host=192.168.1.15:8096

2015-06-08 14:24:27.3437,Info,,Sending Http Get to http://192.168.1.15:8096/mediabrowser/Users/f23a949b-5da2-dd2c-b526-bf4fa8738d39/Items?ParentId=1c03d9463c1bee13f6676676918ce5b8&sortBy=&CollapseBoxSetItems=True&fields=Overview,,Path,,ParentId,,DisplayPreferencesId,,DateCreated,,IndexOptions,,DateLastMediaAdded,,Metascore,,MediaStreams,,SortName,,Taglines,,MediaSources,,SpecialEpisodeNumbers&ImageTypes=&recursive=False&genres=&Ids=&studioIds=&ExcludeItemTypes=&IncludeItemTypes=&ExcludeLocationTypes=&personIds=&personTypes=&years=&format=Json,50,Worker thread for Common

[MrWebsmith 598]

hmm.. welp.. any thoughts on how else i could try to force it.. maybe set them both to some different online time server??

 

im gonna make them both sync with

nist1-pa.ustiming.org

..if i can figure that out.. then ill restest/repost

 

ok.. they are in lock-step to the second..(if you believe widnws) now

 

 

https://www.dropbox.com/s/fayzizm7rncvmrb/timelockstep.jpg?dl=0

 

 

time to retest
.

Edited June 8, 2015 by MrWebsmith

[ebr 14913]

Can you run MBC on the same machine as the server?

[MrWebsmith 598]

sorry man.. no change (to be clear this was the test right after the time sync with nist.gov)

 

mbc: https://www.dropbox.com/s/waiq9w9j2sld68x/MBClassic-8620150c1745c973ed4c91a7b37bc5081ce554.log?dl=0

 

emby server: https://www.dropbox.com/s/v7g0kn2tlhwmtoc/server-63569352156.txt?dl=0

Edited June 8, 2015 by MrWebsmith

[MrWebsmith 598]

Can you run MBC on the same machine as the server?

umm... sure.. give me a sec.. not sure that thing has wmc on it or not.. happy to keep testing though

[paul777 13]

HI, just had a quick review of this thread so apologies if anything I mention has already been said/tried. The changes that were made altered the background threading model to reduce the total number of threads we were creating but at the same time increased the number available for a given operation type (e.g. a server request). What seems to be happening at least in the case at the start of the thread is that the server is giving up when hit will lots of parallel requests and timing out, these requests which timeout in turn cause the common thread pool to block until they do timeout. The current threadpool size is 10 (async.cs ln400) it may be worth testing with a smaller number of threads to see if the server plays better.

 

Also - are the machines on a wireless or wired connection?

 

What would be really great is if you could send me a dump of the process taken during the hang/delay, this will be a process called ehexthost.exe and can be dumped by bringing up task manager, right clicking the process and selecting Create Dump File. The file it creates may well be fairly large. It none of that makes sense to your don't worry. It you are able to do it please confirm which version you are on at the time.

[MrWebsmith 598]

hey paul.. 

 

 

all machines are wired gig connections to server.. just updated to 279 it seems.. i was 278 above

 

im getting mbc quickly setup locally on the server itself for ebr, but i will retest and try to add in the process dump during the delay and post it with the next round of logs

[MrWebsmith 598]

welp.. my scientific method really sucks.. cause now the server "htpc" is a new variable..but so is 279...

 

on 279 on the local server box itself (mbc and server are same windows machine) with the custom view ON... im not getting a delay...

 

im about to retest the normal htpc we've been testing all day with 279 (this would be the only new variable) and keeping the custom tv series view as ON... on 278 there was a delay in this config just a bit ago, so we'll see

[ebr 14913]

Paul - I'm trying to confirm with Web's help but I don't think the server is actually timing out.  If you look at the requests on the server, they are coming in and responding in a fraction of a second.

 

What appears to be happening is that the client side is locking up after some number of simultaneous requests to the server.  I've put in a ton of logging and can see the request sending and then the thread (and everything else) blocks until the specified time-out period.  However, it isn't the server that is timing out.  It appears it is something in the client side.

 

I have just found a setting on WebRequest that I increased in 279.  We'll see if that was the culprit when Web reports back but, based on his experience with localhost, I think maybe it will be as that setting isn't used on a local connection.

[MrWebsmith 598]

279 even with the custom view ON.. is OK

 

on both my normal living rm HTPC... it is OK

 

and on the server as a local mbc install it is also OK

 

 

would either of you like more testing/or confirmation logs?... paul do you want a process dump even if no delay?

[MrWebsmith 598]

279 with custom view OFF.. still OK on the main htpc 

 

 (just rechecked to be paranoid and i tend to mostly NOT have the custom view set