From d4a3e26534334cdbb9111fca3c430b3b2913885a Mon Sep 17 00:00:00 2001
From: Dimitri Savineau <dsavinea@redhat.com>
Date: Tue, 7 May 2019 16:33:21 -0400
Subject: [PATCH] ceph-handler: Fix rgw socket in restart script

Since Mimic the radosgw socket has two extra fields in the socket
name (before the .asok suffix): <pid>.<ctid>

Before:
  /var/run/ceph/ceph-client.rgw.cephaio-1.asok
After:
  /var/run/ceph/ceph-client.rgw.cephaio-1.16913.23928832.asok

The radosgw restart script doesn't handle this and could fail during
an upgrade.
If the SOCKETS variable isn't defined in the script then the test
command won't fail because the return code is 0

$ test -S
$ echo $?
0

There multiple issues in that script:
  - The default SOCKETS value isn't defined due to a typo
SOCKET vs SOCKETS.
  - Because the socket name uses the pid then we need to check the
socket name after the service restart.
  - After restarting the radosgw service we need to wait few seconds
otherwise the socket won't be created.
  - Update the wget parameters because the command is doing a loop.
We now use the same option than curl.
  - The check_rest function doesn't test the radosgw at all due to
a wrong test command (test against a string) and always returns 0.
This needs to use the DOCKER_EXECS variable in order to execute the
command.

$ test 'wget http://192.168.100.11:8080'
$ echo $?
0

Also remove the test based on the ansible_fqdn because we only use
the ansible_hostname + rgw instance name.

Finally group all for loop into a single one.

Resolves: #3926

Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
(cherry picked from commit c90f605b5148d179790cec545d02db1086579994)
---
 .../templates/restart_rgw_daemon.sh.j2        | 29 +++++++------------
 1 file changed, 10 insertions(+), 19 deletions(-)

diff --git a/roles/ceph-handler/templates/restart_rgw_daemon.sh.j2 b/roles/ceph-handler/templates/restart_rgw_daemon.sh.j2
index e78a54102..4f27de050 100644
--- a/roles/ceph-handler/templates/restart_rgw_daemon.sh.j2
+++ b/roles/ceph-handler/templates/restart_rgw_daemon.sh.j2
@@ -14,19 +14,14 @@ fi
 declare -a DOCKER_EXECS
 for ((i=0; i<${RGW_NUMS}; i++)); do
   DOCKER_EXECS[i]=""
-{% if containerized_deployment %}
+{% if containerized_deployment | bool %}
   CONTAINER_NAME="ceph-rgw-${HOST_NAME}-rgw${i}"
   DOCKER_EXECS[i]="{{ container_binary }} exec ${CONTAINER_NAME}"
 {% endif %}
 done
 declare -a SOCKETS
-# Backward compatibility
-for ((i=0; i<${RGW_NUMS}; i++)); do
-  SOCKET[i]="EMPTY_SOCKET"
-  ${DOCKER_EXECS[i]} test -S /var/run/ceph/{{ cluster }}-client.rgw.{{ ansible_fqdn }}.asok && SOCKETS[i]=/var/run/ceph/{{ cluster }}-client.rgw.{{ ansible_fqdn }}.asok
-  ${DOCKER_EXECS[i]} test -S /var/run/ceph/{{ cluster }}-client.rgw.${HOST_NAME}.rgw${i}.asok && SOCKETS[i]=/var/run/ceph/{{ cluster }}-client.rgw.${HOST_NAME}.rgw${i}.asok
-done
 RGW_IP={{ hostvars[inventory_hostname]['_radosgw_address'] }}
+SOCKET_PREFIX="/var/run/ceph/ceph-client.rgw.${HOST_NAME}.rgw"
 
 check_socket() {
   local i=$1
@@ -34,6 +29,7 @@ check_socket() {
   local count=10
   # Wait and ensure the socket exists after restarting the daemon
   while [ $count -ne 0 ]; do
+    ${DOCKER_EXECS[i]} test -S ${SOCKET_PREFIX}${i}.*.*.asok && SOCKETS[i]=$(stat --printf=%n ${SOCKET_PREFIX}${i}.*.*.asok)
     ${DOCKER_EXECS[i]} test -S ${SOCKETS[i]} && succ=$((succ+1)) && break
     sleep $DELAY
     let count=count-1
@@ -48,7 +44,7 @@ check_socket() {
 check_for_curl_or_wget() {
   local i=$1
   if ${DOCKER_EXECS[i]} command -v wget &>/dev/null; then
-    rgw_test_command="wget --quiet"
+    rgw_test_command="wget --tries 1 --quiet -O /dev/null"
   elif ${DOCKER_EXECS[i]} command -v curl &>/dev/null; then
     rgw_test_command="curl --fail --silent --output /dev/null"
   else
@@ -63,28 +59,23 @@ check_rest() {
   check_for_curl_or_wget ${i}
   local succ=0
   while [ $RETRIES -ne 0 ]; do
-    test "$rgw_test_command $RGW_PROTOCOL://$RGW_IP:$((RGW_BASE_PORT+i))" && succ=$((succ+1)) && break
+    ${DOCKER_EXECS[i]} $rgw_test_command $RGW_PROTOCOL://$RGW_IP:$((RGW_BASE_PORT+i)) && succ=$((succ+1)) && break
     sleep $DELAY
     let RETRIES=RETRIES-1
   done
   if [ $succ -ne 1 ]; then
     # If we reach this point, it means there is a problem with the connection to rgw
-    echo "Error connecting locally to Rados Gateway service: $RGW_PROTOCOL://$rgw_listen"
+    echo "Error connecting locally to Rados Gateway service: $RGW_PROTOCOL://$RGW_IP:$((RGW_BASE_PORT+i))"
     exit 1
   fi
 }
 
-# First, restart the daemon
 for ((i=0; i<${RGW_NUMS}; i++)); do
+  SOCKETS[i]="EMPTY_SOCKET"
+  # First, restart the daemon
   systemctl restart ceph-radosgw@rgw.${HOST_NAME}.rgw${i}
-done
-
-# Check socket files
-for ((i=0; i<${RGW_NUMS}; i++)); do
+  # Check socket files
   check_socket ${i}
-done
-
-# Check rest
-for ((i=0; i<${RGW_NUMS}; i++)); do
+  # Check rest
   check_rest ${i}
 done