tools: Configure systemd to always restart FRR, regardless of exit code

The current service file configures restarts on-abnormal, which translates to "unclean signal", "timeout", or "watchdog". This patch updates it to always restart, as there's never really a time watchfrr should exit by itself at all.

Signed-off-by: Brian Rak <brak@vultr.com>

diff --git a/tools/frr.service.in b/tools/frr.service.in
index df1e4f3b081e0f47be669949cf2e7dedb4f2446b..1e958dd93e3d0c7bec4ad809c4822289bf10bab2 100644 (file)
--- a/tools/frr.service.in
+++ b/tools/frr.service.in
@@ -15,7 +15,7 @@ StartLimitBurst=3
 TimeoutSec=@TIMEOUT_MIN@m
 WatchdogSec=60s
 RestartSec=5
-Restart=on-abnormal
+Restart=always
 LimitNOFILE=1024
 PIDFile=@CFG_STATE@/watchfrr.pid
 ExecStart=@CFG_SBIN@/frrinit.sh start

diff --git a/tools/frr@.service.in b/tools/frr@.service.in
index 1cbef1b18c042573ce88c7261eecd52300745dda..85408a0cc7ea99b26b9024246fdb1686faa3d3a0 100644 (file)
--- a/tools/frr@.service.in
+++ b/tools/frr@.service.in
@@ -15,7 +15,7 @@ StartLimitBurst=3
 TimeoutSec=@TIMEOUT_MIN@m
 WatchdogSec=60s
 RestartSec=5
-Restart=on-abnormal
+Restart=always
 LimitNOFILE=1024
 PIDFile=@CFG_STATE@/%I/watchfrr.pid
 ExecStart=@CFG_SBIN@/frrinit.sh start %I