sysrepo: ensure that stuck /run/sysrepo SHM can be trashed
When the sysrepo stack dies, we were trying hard to just shoot all the
services which use that filesystem to restart, unmount that, and start
from scratch. Unfortunately, this was not really happening. When there
was any lingering process that was still accessing the filesystem, such
as `sysrepo-cli`, the unmount would fail.
Fix that by doing the equivalent of passing the `umount` options `-f`
for "force unmount" (which I was thinking would do the job, but nope, it
doesn't), and `-l` (lazy unmount). Together, they do the trick.
Change-Id: Icb0c706d8334f8434df0e4b97cccca893a2c9839
diff --git a/package/reset-sysrepo/run-sysrepo.mount b/package/reset-sysrepo/run-sysrepo.mount
index 26f5a84..61c9eca 100644
--- a/package/reset-sysrepo/run-sysrepo.mount
+++ b/package/reset-sysrepo/run-sysrepo.mount
@@ -7,3 +7,5 @@
Where=/run/sysrepo
Type=tmpfs
DirectoryMode=0000
+LazyUnmount=yes
+ForceUnmount=yes