The issue caused during the upgrade process due to certain steps that rely on Linux tools (like `curl`
) and Python scripts, which do not honor the proxy configuration file located at '/etc/sysconfig/proxy'.
Aria Lifecyle Manager 8.X
During the post-update phase, the script '/var/lib/vrlcm/populate.sh'
checks service status using a 'curl'
command, "curl -I http://localhost:8080/lcm/local/getpassword/serviceadmin@local | head -n 1 | cut -d$' ' -f2"
Despite disabling the proxy with `PROXY_ENABLED="no"
` in '/etc/sysconfig/proxy'
, 'set | grep proxy'
still shows proxy variables in the shell environment, causing 'curl'
to malfunction and incorrectly report services as down.
The root cause seems to be that a script reads values from '/etc/sysconfig/proxy'
but ignores the 'PROXY_ENABLED="no"'
setting, leading to improper management of proxy variables.
Workaround:
The current workaround is to temporarily unset the proxy configuration during the upgrade process.
In the future, an enhancement will allow proxy support during the upgrade process, where a common config file will be available for updating proxy settings.
Once the enhancement is implemented in future releases, the proxy configuration will be supported during upgrades, resolving the issue.
However, until then, the only workaround is to manually unset the proxy settings during the process to ensure 'curl'
functions correctly.
Note:
See the Configure your proxy settings section of the VMware Aria Suite Lifecycle 8.x product documentation.
Also it can be disable by removing the proxy configuration in '/etc/sysconfig/proxy'.