CTIOS errors - pending writes, exceeds limit. – Cisco DevNet Developer Support

In our environment, I discover multiple messages (related to one recorder) on CTIOS queue, they looks like cloned or multipled, example:

08:47:18:627 CTIOS1-ctios Trace: <-SND[07096][agent.5000.29327]:Monitored eAgentStateEvent(eTalking),Msk(0)

08:47:18:643 CTIOS1-ctios Trace: <-SND[07096][agent.5000.29327]:Monitored eAgentStateEvent(eTalking),Msk(0)

which leading to error:

08:47:18:549 CTIOS1-ctios Trace: IOCPConnectionMgrThread[6060],Client[07264, 05CE0600]:Pending Writes(8000) Exceeds Limit(8000)Client disconnected by CTIOS Server,Client(7264) at Address(X.X.X.X),Total Client pending write operations is (14464) for (121) Client(s) connected,LastWriteResponseTime(0 ms)

Assuming the queue exceeds more than 8000 pending writes, and then CTIOS treat the client as a slow client and closes the connection, I don't known what's the reason of this situation, and which cause may result in such behavior?

Do you have, any clue why queue exceeded pending writes (8000), and what can cause this (maybe wrong configuration of monitor mode's)?

I don't think it's normal behaviour, i mean whether there should be much less messages in that short range of time.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Appolagies for the delay in answering to the quriey. Following is the answer i got from the CTIOS Team.

The Client disconnect event usually means that client is very slow in processing the messages and server has a buffer in which it holds messages. If buffer is exceeded server cannot accept anymore messages for that client. If queue exceeds more than 12000 pending writes then CTIOS treat the client as a slow client and closes the connection so that performance of CTIOS will not degrade.

To summarize here, in this scenario this could be of two suspects:

1. Client machine itself is slow to process the messages coming from server.

2. Agent facing the issue because of more than 50 ( As per the defined limits ) skill groups assigned for the problematic agent

For case 2 if customer have the faulty agent with more than 50 skill group assigned , then they must have to reduce the assigned skillgoups count and monitor the system. Request to verify.

Better to verify the environment according to sizing limits documented by cisco.

(Table 2 Sizing Effects Due to Number of Skill Groups/Precision Queues per Agent (12,000 Agents)

Ref: http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cust_contact/contact_center/ipcc_enterprise/ippcenterprise9_0_1/design/guide/UCCE_BK_S06086EC_00_srnd-9-0-1/UCCE_BK_S06086EC_00_srnd-9-0-1_chapter_01001.html

please ensure that you running the setup under the limit of Maximum CTI connections as documented by cisco

Please find below summary regarding the limits as documented in cisco docs :

1. Total number Monitor mode connections to CTIOS alone - 5.

2. Total number of All Event clients to CTIServer alone - 7.

3. Total number of Monitor mode and All Event clients to both CTIOS Server and CTIServer combined together - 9.

Please collect the procmon output for clients command:

You need to use “clients” command from procmon utility to validate the connections.

Here is an example for the same:

C:\icm\serviceability\wsccli>procmon icm10 cg1a ctisvr

>>>>clients

Session Time Ver Flags ClientID AgentID AgentExt Signature Host

2 74 days 15 AUX CTIOSServer CTIOSServer (10.86.138.153:64163)

since we have only one connection setup, only one session listed . For every new connection, you can locate more rows listed

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Articles in this section

CTIOS errors - pending writes, exceeds limit.

Comments