chore: ensure bash wraper in emqx.service, restart wait 60 seconds

CHANGES-4.3.md

@@ -26,6 +26,7 @@ File format:
 * `topic` parameter in bridge configuration can have `${node}` substitution (just like in `clientid` parameter)
 * Add UTF-8 string validity check in `strict_mode` for MQTT packet.
   When set to true, invalid UTF-8 strings will cause the client to be disconnected. i.e. client ID, topic name. [#7261]
+* Changed systemd service restart delay from 10 seconds to 60 seconds.
 
 ### Bug fixes
 
@@ -40,6 +41,7 @@ File format:
 * Fix rpc get node info maybe crash when other nodes is not ready.
 * Fix false alert level log “cannot_find_plugins” caused by duplicate plugin names in `loaded_plugins` files.
 * Prompt user how to change the dashboard's initial default password when emqx start.
+* Fix errno=13 'Permission denied' Cannot create FIFO boot error in Amazon Linux 2022 (el8 package)
 
 ## v4.3.12
 ### Important changes

deploy/packages/emqx.service

@@ -7,11 +7,18 @@ User=emqx
 Group=emqx
 Type=forking
 Environment=HOME=/var/lib/emqx
-ExecStart=/usr/bin/emqx start
+
+# Must use a 'bash' wrap for some OS
+# errno=13 'Permission denied'
+# Cannot create FIFO ... for writing
+ExecStart=bash /usr/bin/emqx start
+
 LimitNOFILE=1048576
-ExecStop=/usr/bin/emqx stop
+ExecStop=bash /usr/bin/emqx stop
 Restart=on-failure
-RestartSec=5s
+
+# When clustered, give the peers enough time to get this node's 'DOWN' event
+RestartSec=60s
 
 [Install]
 WantedBy=multi-user.target

deploy/packages/rpm/emqx.service

@@ -1,17 +0,0 @@
-[Unit]
-Description=emqx daemon
-After=network.target
-
-[Service]
-User=emqx
-Group=emqx
-Type=forking
-Environment=HOME=/var/lib/emqx
-ExecStart=/usr/bin/emqx start
-LimitNOFILE=1048576
-ExecStop=/usr/bin/emqx stop
-Restart=on-failure
-RestartSec=5s
-
-[Install]
-WantedBy=multi-user.target

deploy/packages/rpm/emqx.service

@@ -0,0 +1 @@
+../emqx.service