openstack/tripleo-common
Code Issues Proposed changes

By default, unset everything proxy related for healthcheck

Browse Source 
Hopefully get out of the hell of proxy/no_proxy settings for
healthchecks.

Since operators continue to (ab)use the http_proxy, https_proxy and,
especially, the no_proxy environment variables, and since this has been
creating many, many, many issues within container healthchecks, and,
finally, since proxy settings are mostly useless for that specific
usecase, we just discard them by default.

With this patch, the different healthchecks will go without using any
proxy. This is mostly possible because checks are done on "internal"
networks provided by TripleO, and there's no proxy in-between. Most of
the connections are even on the local node, on its different and various
associated IPs, so there's, really, no need to obey any proxy setting.

Change-Id: Ic7928f7895d10aa4a8eb1d4e1f55fd30de83aee4
Resolves: rhbz#2121349
This commit is contained in:
Cédric Jeanneret 2022-10-04 10:18:18 +02:00
parent e310a3ced0
commit ec3e9fc276
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
healthcheck/common.sh
View File

@ -12,6 +12,13 @@ fi
: ${HEALTHCHECK_CURL_PY_USER_AGENT:=pyrequests-healthcheck}
: ${HEALTHCHECK_CURL_WRITE_OUT:='\n%{http_code} %{remote_ip}:%{remote_port} %{time_total} seconds\n'}
: ${HEALTHCHECK_CURL_OUTPUT:='/dev/null'}
: ${HEALTHCHECK_USE_PROXY:=0}
if [ $HEALTHCHECK_USE_PROXY -eq 0 ]; then
for prox_var in http_proxy https_proxy no_proxy; do
unset $prox_var
unset ${prox_var^^}
done
fi
get_user_from_process() {
process=$1