Skip to content

Commit 7e0d63a

Browse files
authored
Merge pull request #1595 from NVIDIA/backport-1586-to-release-1.18
[release-1.18] fix: Allow nvidia-cdi-refresh.service to be restarted
2 parents 38d26bf + 7538b5e commit 7e0d63a

1 file changed

Lines changed: 7 additions & 0 deletions

File tree

deployments/systemd/nvidia-cdi-refresh.service

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -17,6 +17,9 @@ Description=Refresh NVIDIA CDI specification file
1717
ConditionPathExists=|/usr/bin/nvidia-smi
1818
ConditionPathExists=|/usr/sbin/nvidia-smi
1919
ConditionPathExists=/usr/bin/nvidia-ctk
20+
# Limit the number of successive restarts to 5 in 10 seconds.
21+
StartLimitBurst=5
22+
StartLimitIntervalSec=10s
2023

2124
[Service]
2225
Type=oneshot
@@ -26,6 +29,10 @@ EnvironmentFile=-/etc/nvidia-container-toolkit/nvidia-cdi-refresh.env
2629
ExecCondition=/usr/bin/grep -qE '/(nvidia|nvidia-current)\\.ko' /lib/modules/%v/modules.dep
2730
ExecStart=/usr/bin/nvidia-ctk cdi generate
2831
CapabilityBoundingSet=CAP_SYS_MODULE CAP_SYS_ADMIN CAP_MKNOD
32+
# We set the service to restart on failure to ensure that a CDI spec is
33+
# eventually generated.
34+
Restart=on-failure
35+
RestartSec=1s
2936

3037
[Install]
3138
WantedBy=multi-user.target

0 commit comments

Comments
 (0)