RTSP doesnt answer every time (freezes)

Managing an App-V environment; Check-lists, Tips and Tricks, Best Practices, Videos, How to Guides

Moderators: kirk, jur, kkaminsk

Post Reply
verdict
Still installing locally
Posts: 2
Joined: Thu Feb 16, 2012 10:24 am

RTSP doesnt answer every time (freezes)

Post by verdict » Thu Feb 16, 2012 11:17 am

Hi,

we are currently running App-V 4.6, Windows 2008R2. We have to separate management servers and everything have worked kinda well earlier. We have NetApp Shared RO cache (working nowdays :D), and SQL Mirroring. We also do NLB with Cisco for App-V streaming servers. SFTFS.SFT cahcefile is currenly a bit over 10 GB.

We have start having a problems with our environment. Both management servers seems to got some weird problems, RTSP is not answering always. Clients got several different errors, and management servers got several errors in event logs. Both servers are quite fresh, they shouldn't have any hardware issues (they are running @ VMWare ESX, 4 gig RAM, 2 cores).

When I check connection from Cisco (I've created probe that asks server via rtsp), they seems to go down and stay down like 1 to 60 minutes. Totally random downtime. They come up without any intervention, but also boot fixes problem right away. They may go down same time, or alone. They may also work like 6 hours straight without any problems, or go down in one minute, get up in next and go down again in next.

Servers have these errors in event logs, and they have appeared during these couple last days that we are encountering these problems:

41495 - Server record not found in data source for following: SERVERNAME <-- many of these
41485 - File Cache is full

Also single instances of these..

41505 - Failed to initialize RTSPS
41472 - Module execution failed

Client got errors:

"No connection could be made because the target machine actively refused it"
"A network operation did not complete in time .. "

As resolution I've tried:
1. Set filecache from 512MB to 2048MB. This should not be problem, because we only stream to servers, and there are only like 50 of them.
2. Monitoring servers memory and CPU usage. Both are fine.
3. Check and doublecheck DHCP and DNS. Actually it cant be these, because everything was working earlier like a swiss swatch. Whole AD and environment is only couple months old. No changes were made when these problems started.
4. Telnet during "blackout" to SERVERNAME01 322 and 554 and they answer in this level.
5. Check Wireshark logs during blackout. Nothing special as far as I know.

Questions:
1. What "Database Size" configures? What are stored in there? I've set this from 1024 to 2048 (rightclick SERVERNAME and choose System options).
2. What role our mirrored SQL plays? Our main SQL admin is @ vacation :/
3. How I can define that RTSP is not answering? Only way to see it is via Cisco probe. Server responds and everythings seems to be in order, but RTSP is not answering. Also

kirk
Guru
Posts: 810
Joined: Fri Aug 27, 2004 3:32 pm
Location: Germany
Contact:

Re: RTSP doesnt answer every time (freezes)

Post by kirk » Thu Feb 16, 2012 2:36 pm

Hi,

the log extracts you sent indicate communication issues with the database, but that's not necessarily the root cause. However, I'd start there:
Can both Servers access the database, what'S happening when the SQL Principal goes down and the connections should switch to the Mirror (probably you can't test this without your DB Admin).

Verify the steps outlined at http://technet.microsoft.com/en-us/libr ... 60790.aspx

There is an MS KB Articel about the Server Error Codes: http://support.microsoft.com/kb/931165/en-us but I asume this isn't the issue with your deployment (because the article describes a static, initial issue)
Just make sure that in the App-V Management Console / Server Groups your App-V Management Servers are listed and that both machine accounts do have access to the SQL database.
In fact I asume this is propperly configured as well (otherwise you'd see issues earlier).

An obvious question is if anything changed recently. Or did the amount of users increase in the past (or did you add applications that are more frequently used)?


- The DB size of 1-2 GB is appropriate for most environments. The "important" configuration and publishing information consume a quite samll amount ot it anyway. Most is reporting data (App Launch and Shutdowns)

- I strongly recommend that you check the "Max Memory Allocation" value by the App-V Management Servers (Server MMC / Server Groups / <Group-name>: Properties of each individual server. It is set to 512MB by default and should be increased to a reasonable" amount (= Machine_RAM - 1GB_for_OS), so 6-7 GB on a 8 GB box. http://blog.zarb.co.uk/2011/08/app-v-ti ... emory.html talks about that. This settings requires a service restart ofg the App-V Management Server service on every node. Actually, this is my most preferred recommendation here.

- For RTSP monitoring, you could use your LoadBalancer (sophistcated solutions allow for "well-formatted-answers" checks). As a Windows solution, I recommend immidio's "App-V Ping Tool" http://immidio.com/resourcekit/. It's free and can be automated.
Falko

Post Reply

Return to “Managing”

Who is online

Users browsing this forum: No registered users and 1 guest