IBM Support

IBM Cloud Pak System (or IBM PureApplication System): NTP time not synchronizing correctly on AIX virtual machine

Troubleshooting


Problem

NTP time on AIX 6.x, 7.x virtual machine(s) are not synchronized with the NTP time of the System's Management Node

Symptom

The Virtual System Instance is stuck in LAUNCHING state after restarting the instance.  The following curl error message is reported in the virtual machine's initialization script log (/0config/0config.log) 
update status to INITIALIZING
CURL_SSL_OPTION = --cacert /0config/cert.tmp/cert_1.pem
updateURL: inputURL = https://[<ipv6>]:9443/services/deployments/<deployment-ID>/command
updateURL: newURL = https://IBMWorkloadDeployer:9443/services/deployments/<deployment-ID>/command
newURL = https://IBMWorkloadDeployer:9443/services/deployments/<deployment-ID>/command
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:--  0:01:15 --:--:--     0
100    95    0     0  100    95      0      1  0:01:35  0:01:15  0:00:20  1283
curl: (22) The requested URL returned error: 401
Failed to update status:  22
 

Diagnosing The Problem

Go to System > System Settings, then expand the "Date and Time" section.   Compare the time here with the time on the virtual machine node(s).  If the times do NOT match then proceed with the suggested steps in the Resolving The Problem section to sync the time and restart the instance.

Resolving The Problem

1. confirm xntpd process is started and running on the VM(s)

#startsrc -s xntpd
#ps -ef | grep xntpd

2. obtain the ntp server address from the /etc/ntp.conf and confirm the address is reachable
ntp server address = This should be the IPv6 address of System's Management Node
#cat /etc/ntp.conf  
e.g.
server <ntp_server_IPv6_address>
#ping <ntp_server_IPv6_address>
#ntpq -p

3. sync the time using ntpdate command
#ntpdate -d <ntp_server_IPv6_address>

Note:  If you see an IPv4 address returned when running ntpdate command...
e.g.
# ntpdate -d fd8c:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx                  
31 Jul 21:55:50 ntpdate[12748543]: 3.4y                             
transmit(253.xxx.xx.xx)                                             
transmit(253.xxx.xx.xx)                                             
transmit(253.xxx.xx.xx)                                             
transmit(253.xxx.xx.xx)                                             
transmit(253.xxx.xx.xx)                                             
server 253.xxx.xx.xx, port 123                                      
...
31 Jul 21:55:50 ntpdate[12748543]: no server suitable for synchronization found      
This is known issue:  
============================
Symptom:
the  IPv6 address stored in /etc/ntp.conf is converted by XNTPD to an IPv4 address, which it cannot find.  XNTPD is taking the first 32-bits of the IPv6 address defined in /etc/ntp.conf and converting it to an IPv4 address, then truncating the rest of the IPv6 bits.

Cause:
VM is using a XNTPD version 3 that is not capable of running in IPv6 mode.  The AIX OS must use ntp4 version which supports IPv6

Solution:
most OS have both version 3 and version 4 installed, so just a matter for changing the symlinks from ntp3 --> ntp4

see AIX doc for instructions
https://www-01.ibm.com/support/docview.wss?uid=ibm10738319 
=============================
4. restart the Virtual System Instance to bring the instance back to RUNNING state.     
                          

Document Location

Worldwide

[{"Business Unit":{"code":"BU053","label":"Cloud & Data Platform"},"Product":{"code":"SSFQSV","label":"IBM Cloud Pak System Software"},"Component":"","Platform":[{"code":"PF002","label":"AIX"}],"Version":"All Versions","Edition":"","Line of Business":{"code":"LOB45","label":"Automation"}}]

Document Information

Modified date:
06 May 2020

UID

ibm11005640

Page Feedback