Processes ccirmtd and cciclnd utilizing more than 50% CPU

Document ID : KB000008481
Last Modified Date : 14/02/2018
Show Technical Document Details
Issue:

In the Unix syslog file, there are several messages where CCI is trying to establish a connection with a remote IP address

daemon:info CCIRMTD[28377236]: .CAICCI_W_2014[xxxxxxxxx] Attempt to connect to host < IP Address> failed, Check to see if Unicenter is active on < IP Address>

daemon:info CCIRMTD[28377236]: .CAICCI_I_2051 errno=<57>, Unable to connect to remote host < IP Address>

daemon:info CCIRMTD[28377236]: .CAICCI_W_2014[xxxxxxxxx] Attempt to connect to host < IP Address> failed, Check to see if Unicenter is active on < IP Address>

Environment:
UUJMA 11.2 SP2 on Linux but it can happen with any other Unix platform
Cause:

As soon as any CCI connection is registered in the CCI Host Address Table from any target machine, CCI will try to connect to this node.

If for whatever reason this node is no longer there, these messages will be written to the Unix Syslog file.

We have also seen such a behavior when a port scanner reaches out to the CAICCI port number, which is default port number 1721.

All these CCI retries are consuming a large number of CPU cycles.

Resolution:

To fix such a problem, you can set a CCI variable to prevent all these CCI retries among nodes which are no longer accessible and which are not defined in the ccirmtd.prf file.

1) Create this below envsetlocal file with the following statements :

vi $CAIGLBL0000/cci/scripts/envsetlocal

CAI_CCI_NORECON=Y

export CAI_CCI_NORECON

2) When done, you'll have to source the UUJMA environment, or run : su - root

3) Then check if the variable is properly set with: "env |grep CAI_CCI_NORECON"

4) Then restart UUJMA