[ARVADOS] updated: 2.1.0-1876-g12c1c5131

[Git user]

Summary of changes:
 tools/compute-images/scripts/base.sh | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

       via  12c1c51313e897abd0e9d1801b42bc8dc3b8d1d9 (commit)
      from  ac52d7ee23b39779712c702945eb9db7e17dd814 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.


commit 12c1c51313e897abd0e9d1801b42bc8dc3b8d1d9
Author: Ward Vandewege <ward at curii.com>
Date:   Thu Feb 3 15:42:18 2022 -0500

    18713: expand comment.
    
    Arvados-DCO-1.1-Signed-off-by: Ward Vandewege <ward at curii.com>

diff --git a/tools/compute-images/scripts/base.sh b/tools/compute-images/scripts/base.sh
index b30ad38d1..450a8b3c5 100644
--- a/tools/compute-images/scripts/base.sh
+++ b/tools/compute-images/scripts/base.sh
@@ -210,7 +210,8 @@ if [ "$NVIDIA_GPU_SUPPORT" == "1" ]; then
   # This service fails to start when the image is booted without Nvidia GPUs present, which makes
   # `systemctl is-system-running` respond with "degraded" and since that command is our default
   # BootProbeCommand, compute nodes never finish booting from Arvados' perspective.
-  # Disable the service to avoid this.
+  # Disable the service to avoid this. This should be fine because crunch-run does its own basic
+  # CUDA initialization.
   $SUDO systemctl disable nvidia-persistenced.service
 fi
 

-----------------------------------------------------------------------


hooks/post-receive
--