How to fix clstat and cldump related problems

How to fix clstat and cldump related problems

A common problem with clstat and cldump commands is often observed especially when a new PowerHA (HACMP) cluster is configured on a freshly installed AIX. Most of the issues with clstat and cldump are related to snmp protocol and can be solved with simple correction of the snmp configuration file.

Follow the procedure below if clstat and cldump are not working:

1. Check which version of snmp is used:

# ls -l /usr/sbin/snmpd

The following options are possible:

 snmpdv1 - snmpd V1 only daemon

 snmpd64v1 - snmpd V1 only 64 bit daemon

 snmpdv3ne - snmpd V3/V2/V1 non-encrypted daemon

 snmpdv3e - snmpd V3/V2/V1 encrypted daemon

By default, /usr/sbin/snmpd is linked to snmpdv3ne.

In order to change to different version use the following command and keys:

# /usr/sbin/snmpv3_ssw -n =>back to snmpdv3ne

# /usr/sbin/snmpv3_ssw -e =>back to snmpdv3e

# /usr/sbin/snmpv3_ssw -1 =>to snmpdv1

2. If snmpdv3 is used the easiest way is to switch to snmpdv1:

#  stopsrc -s clinfoES

#  stopsrc -s snmpd

#  /usr/sbin/snmpv3_ssw -1 (or switching to version 3 of the protocol)

#  startsrc -s snmpd

#  startsrc -s clinfoES

However you can fix snmpdv3 problems with the steps below (recommended).

3. If clstat and cldump are not working, check /etc/snmpdv3.conf for the following:    

- Does the following line exist, if not add it:                                           

VACM_VIEW defaultView        1.3.6.1.4.1.2.3.1.2.1.5    - included - 

(this line will allow access to "risc6000clsmuxpd" but not to complete "internet" MIB)
 
Alternatively you can uncomment the line:

VACM_VIEW defaultView    internet       - included - )

- Remove the comments in the following lines (especially for AIX6.1, AIx 7.1, IZ41204):                       

   smux     1.3.6.1.4.1.2.3.1.2.1.2         gated_password # gated   

   smux     1.3.6.1.4.1.2.3.1.2.1.5      clsmuxpd_password # PowerHA  SystemMirror clsmuxpd                                                 

to look like that:                                                   

   smux     1.3.6.1.4.1.2.3.1.2.1.2         gated_password           

   smux     1.3.6.1.4.1.2.3.1.2.1.5      clsmuxpd_password           

- Check if comment exists in front of the following line, if it does, remove it:

#COMMUNITY public    public     noAuthNoPriv 0.0.0.0   0.0.0.0             -

(For security reason it is recommended to change the word public to something else)

4. Restart deamons and services:

# stopsrc -s clinfoES

# stopsrc -s snmpd

do the change

#  startsrc -s snmpd

#  startsrc -s clinfoES

Check:

wait 1-2 minutes

#  snmpinfo -m dump -v -o /usr/es/sbin/cluster/hacmp.defs  cluster

#  cldump

#  clstat

Or you can restart all the the deamons (recommended if not a production system):

#  stopsrc -s clinfoES                                                 

#  stopsrc -s aixmibd                                                  

#  stopsrc -s snmpmibd                                                 

#  stopsrc -s hostmibd                                                 

#  stopsrc -s snmpd                                                    

 

#  startsrc -s snmpd                                                    

                                                                       

#  tail -f /var/hacmp/log/clstrmgr.debug (wait for the following line: 

"smux_simple_open ok, try smux_register()")                            

 

#  startsrc -s hostmibd                                                

#  startsrc -s snmpmibd                                                

#  startsrc -s aixmibd                                                 

#  startsrc -s clinfoES

#  refresh -s clstrmgrES                                                


tags: cldump, clstat, snmpd


Comments:

Simm

2013-06-18 00:45:50
Hi, thanks for the tips on what part of mib needs to be added in case of snmpdv3.

CG

2013-07-11 05:09:23
Thanks

wou

2015-04-08 04:19:59
it works fine only after launching "cldare" command on both nodes (PowerHA 7.1.3) : /usr/es/sbin/cluster/utilities/cldare -rt -V 'normal' -C'interactive'

EricC.

2016-01-14 12:04:08
For our site, after upgrading a 6.1 cluster to 7.1.3.4, we had to remove the ip6 loopback ::1 entry in /usr/es/sbin/cluster/etc/clhosts. After we did that: stopsrc -s clinfoES refresh -s clstrmgrES startsrc -s clinfoES and clstat started working again reliably.

Leave a Comment: