Link down every 2-3 weeks [running CTI HA]

[noor_372]

Hi Guys,

We are running genesys CTI HA , we are facing issues that every 2-3 weeks link goes down.

Can anyone help me what exactly may be the root cause for this issue
I am uploading issue logs.

[tony]

"CTI" is a bit non-specific...  Which component is failing over to HA?

As a pure guess and only if it is the exact same time period for each occurrence, you will probably be looking at some sort of automated maintenance schedule/job on your Databases (Maintenance Schedule), your OS (Windows - AT's, Solaris - CRON) or your Network (Router refreshes)...

Tony

[Adam G.]

Tony, I'm guessing its the PBX link that crashes. Which PBX is anyone's guess.

[Noor]

Its TServer Primary or Backup which goes down every 2-3 weeks.

Pbx is avaya comm mgr s87xx.

I have uploaded the logs as well.

[tony]

Where have you uploaded the logs... 

[kip1]

[quote author=noor_372 link=topic=4023.msg17692#msg17692 date=1238095401]
Hi Guys,

We are running genesys CTI HA , we are facing issues that every 2-3 weeks link goes down.

Can anyone help me what exactly may be the root cause for this issue
I am uploading issue logs.
[/quote]

Give us some hints
Are the Links = auto negotiate? (PBX settings)
Do you have time outs when displaying attach data - The voice part of the call comes properly, but the logical is wrong?

[noor_372]

Where can i upload log file.
=========

@15:16:33.3900 [<<] 00 05 00 00
TP_ErrorNotification
Cause: No reply to heartbeat
15:16:33.390 Std 31014 Error Notification: Link[1], reason [No reply to heartbeat]
@03/16/09 15:16:33.3900 [31014] Error Notification: Link[1], reason [No reply to heartbeat]
@15:16:33.3900 [tsgctm] Link [1,s4] stop
@15:16:33.3900 [tsgctm] LinkReg [1, act=1] clearAll
@15:16:33.3900 [tsgctm] Link [1,s4] closeSession
@15:16:33.3900 [tsgctm] Link [1,s4] Changing state to 5
@15:16:33.3900 [asai] Link [1,s5] (sysReqSessionClose)
TP_DisconnectNotification
@15:16:33.3900 [>>] 04 00 00 00
@15:16:33.3900 [tsgctm] Link [1,s5] Changing state to 0
(link_stop) link [1] 'link-tcp' stop
(tcp_stop) closed on [404]
@15:16:33.3900 [tsgctm] PRIMARY link 1 cleared
@15:16:33.3900 [gctmi] TMsg [EventLinkDisconnected()] distributing to model
@15:16:33.3900 [gctmi] Switch [sCON] distributing EventLinkDisconnected
@15:16:33.3900 [gctmi] Switch [sCON] processLinkDisconnected
@15:16:33.3900 [0] 7.6.007.02 send_to_all: message EventLinkDisconnected

[cavagnaro]

Well seems to be that there is a network problem. Have you talked to your network guys? Discarded traffic and priorities? VLAN aisolated?

[bcyk]

mostly network issue..............

[quote]
@15:16:33.3900 [<<] 00 05 00 00
  TP_ErrorNotification
  [b]Cause: [color=red]No reply to heartbeat[/color][/b]
15:16:33.390 Std 31014 Error Notification: Link[1], reason [No reply to heartbeat]
@03/16/09 15:16:33.3900 [31014] Error Notification: Link[1], reason [No reply to heartbeat]
[/quote]


Please check following items:
  1. 'standard' CTI-server network adapters configuration
      - a dedicated/isolate Ethernet port to AES, the CTI-link
      - another Ethernet port to LAN

  2. debug log files are NOT dumped to database table!

  3. When specify disk file name for debug log files, never use network shared drive even to the local machine
      e..g, \\mytservre\gctilog\tserverdeb.log

[noor_372]

Hi Guys,

Thanks for the feedback.

It looks to us a network problem but can you please let me know how can i prove it easily .

As customer is denying any network issue and since it reoccur after every 3-4 weeks or more time.
May be some network glitch occur on the day when link goes down.

Please help me i need to prove it although i believe its a network but not able to prove it.

I have asked customer to start monitoring their network between CTI & AES.

Any network utility which may help me to prove it a network issue.

[cavagnaro]

Not a task for us, but for network guys, put a sniffer, a ping logger, traces, etc. We as PBX or Genesys can't do much, but network guys should be able to help.

[rapjaw08]

Hi,

  We have a similar issue we are using a MAPD card. We also noticed that upon login in to the MAPD we have an error msg saying Application port is not assigned. Also we whenever the tserver fails we can;t ping the mapd nor access it until we hard reset the MAPD card. Any recommendations