VMware Horizon Community
Dar88
Enthusiast
Enthusiast

vmware view connection broker stop responding on port 443

Hi

time to time (mostly 4hr a part ) my connection broker port 443 goes down and it take 3min to come back up , 

at the time it happened  i see this errors on View log 

2015-01-26T15:44:32.712-05:00 WARN  (116C-0BDC) <HTTPS Connection Processor> [PooledProcessor] Connection processor pool full - new connections may not be handled quickly

did any one else have such issue before ?

more info for the setup :

We are using View 5.3.1 and we have 2 connection broker  (active  and passive setup ) , for health check and load balancing we using  F5 

F5 version is 11.4.1

thank you for help

Reply
0 Kudos
17 Replies
Linjo
Leadership
Leadership

How much memory have you for the connection broker? Recommended is 10GB

// Linjo

Best regards, Linjo Please follow me on twitter: @viewgeek If you find this information useful, please award points for "correct" or "helpful".
Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

each server have 12GB RAM and 4CPU

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

base of my Netstat report for past 24Hr at the time of hiccup the connection on port 443  jump like 50 to 200 then goes back down , do you think some one attacking the port ?

Reply
0 Kudos
dnxv5
Contributor
Contributor

Was this resolved? What was the issue and resolution?

Thanks.

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

No is not , and no one has any idea why is happening , but we still looking

Do you have the same issue ?

Reply
0 Kudos
Zwelgje1976
Contributor
Contributor

We are having te same issue. And we dont have that much sessions open (about 20 connected). It's the second time this happened in about 2 months. I hope someone finds the resolution..

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

Interesting ours start around that time also, but it seem yours issue is better then the one we have but in other hand you have 20 connection and we have 400 +

Can you tell me about your setup?  Want compare it  to ours,

 

We are using windows 2008 R2 with 12GB RAM and 4CPU, the connection broker
are VM and running on ESXI 5.5 and we are using F5 load balancer,

Reply
0 Kudos
Zwelgje1976
Contributor
Contributor

We just had the same issue for the second time today :smileyalert:

We dont have a loadbalancer (yet), instead we use the security server on which the error occurs.

We are unning the same specs (VM and ESXI), but i increased the memory today from 4GB to 10GB. Hope this helps..

I suspect a user doing something wich makes the CPU go crazy, but i have to investigate further. Will post when i find out more...

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

are your external user using any type of VPN , if they do what company ?

do you see the errore message in log ? or  you ony see the port get close ?

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

it seem is not F5  we remove it and still was happening

but i find work around , it seem if i disable "HTTP(S) secure tunnel " and reboot the server  the problem doesn't happens anymore but then we have two issue .

teh work aroudn only work for internal if i do the same for extrnal connection broker  to RDP access will not work

and when you disable this the HTML5 access will not work as well

Reply
0 Kudos
FelixYan
VMware Employee
VMware Employee

Hi Dar 88,

What's your View Client version?

We fixed a similar connection flood problem in View Client side which is related with HTTP(s) secure tunnel. Would you please try to upgrade all your View Client to 3.4 or later(3.5 is available now)? With the new Client version, you can enable your secure tunnel again, then the RDP access from the external network could be work fine.

Thanks for reporting the issue!

Regards,

Felix

Reply
0 Kudos
Le0Getz
Contributor
Contributor

We are seeing the Same thing.  10GB 4 CPU Windows 2008 R2 down for about 10 mins. Then just comes up. We are unable to connect to our URL.  Existing users continue to work fine.  Our VM is not showing any High CPU, MEM or Disk latency.  No warnings or alerts in the Windows logs.  But the security log has the <HTTPS Connection Processor>[PooledProcessor] Connection processor full - new connections may not be handled quickly

VMware View 5.2

Clients range from 2.3 to 3.4

Access is from the Internet only

Reply
0 Kudos
WilliamReid
Enthusiast
Enthusiast

How often is the F5 polling the ports on the Connection Server / Security server?

Reply
0 Kudos
cmoon
Contributor
Contributor

Hello guys, same thing here.

About every 2 days, connection manger stop responding for 5 to 10 minutes, clients receive a timeout when try to connect.

We are using Horizon View 6.0.2. On the conn mgr log, we can see the following MSGs:

2015-10-05T14:27:27.098Z WARN  (0C0C-10A0) <HTTPS Connection Processor> [PooledProcessor] Connection processor pool full - new connections may not be handled quickly

2015-10-05T14:28:02.518Z INFO  (0AC0-0E34) <DesktopControlJMS> [Audit] DISCONNECTED:Server:cn=bd6f159a-0041-4e70-8353-67ad2e8b3972,ou=servers,dc=vdi,dc=vmware,dc=int;Pool:cn=naw-scl-id,ou=server groups,dc=vdi,dc=vmware,dc=int;DNS:NAW-SCL-006.ord.aeropost.local;IP:192.168.103.115;USER:ORD\apscl17;USERDN:cn=s-1-5-21-1947257286-471255681-3969905885-3000,cn=foreignsecurityprincipals,dc=vdi,dc=vmware,dc=int;BROKERUSERSID:S-1-5-21-1947257286-471255681-3969905885-3000;

2015-10-05T14:28:02.518Z INFO  (0AC0-0E34) <DesktopControlJMS> [DesktopTracker] User ORD\apscl17 disconnected from machine NAW-SCL-006 for desktop naw-scl-id - session allocated at October 5, 2015 12:51:15 PM UTC, connected for 96 mins 47 secs

2015-10-05T14:28:25.881Z WARN  (0C0C-1184) <pool-1-thread-5> [PooledProcessor] Problem processing HTTP connection: Read timed out

2015-10-05T14:28:25.882Z WARN  (0C0C-12AC) <pool-1-thread-7> [PooledProcessor] Problem processing HTTP connection: Read timed out

2015-10-05T14:28:25.883Z WARN  (0C0C-13E8) <pool-1-thread-14> [PooledProcessor] Problem processing HTTP connection: Read timed out

2015-10-05T14:28:25.883Z WARN  (0C0C-10A8) <pool-1-thread-2> [PooledProcessor] Problem processing HTTP connection: Read timed out

Any help to resolve this?

Thanks and regards, Carlos Luna

Reply
0 Kudos
doernjoe
Contributor
Contributor

Hi all,

same issue here. Appeared on 5.3.1 over the last week, about once a day. After upgrading to 5.3.5 this now occurs about two to four times per day.

If this happens we see ws_tunnelservice.exe ("VMware View Security Gateway Component" service, wstunnel) at 100% CPU (taking one core/vCPU entirely). Workaround is to restart the service.

Anybody got any idea on how to fix this?

Thanks in advance.

Reply
0 Kudos
Dar88
Enthusiast
Enthusiast

Vmware saying they fix this issue with View 6.1.1 and Client 3.4 or newer

the problem is you don't know who using what version of Client unless some one knows how you can do that ??!! which making the fix hard

We already upgrade the server and but still the problem show up .We are send notification to our user to upgrade the client but then I can test it again until I make sure people upgrade their client which making not easy  . For now I am using the work around which I find but you can only use it for internal connection for external user which use security server if you using that you will lose the  RDP access

Work around is to “disable “Secure tunnel connection “if you do that on 5.* you will lose the HTML5 as well but on 6 they separate them.

Reply
0 Kudos
respodu
Enthusiast
Enthusiast

We re also having this issue on 6.1.1 at once a day. Our F5 sends us alerts about the server being down.

Reply
0 Kudos