-
Bug
-
Resolution: Done
-
Highest
-
None
-
None
-
OOM-El Alto Sprint 1
version: casablanca
page: submodules/oom.git/docs/oom_setup_kubernetes_rancher
The "slave_nfs_node.sh" script found at: https://onap.readthedocs.io/en/casablanca/submodules/oom.git/docs/oom_setup_kubernetes_rancher.html#onap-on-kubernetes-with-rancher, currently obtainable by link with `wget https://onap.readthedocs.io/en/casablanca/_downloads/91443f8e0ad897ea9d7fb384edd29153/slave_nfs_node.sh` has a non-breaking UTF-8 space (hex 'C2 A0') in the final line of the script that gets appended to /etc/fstab. Even though the script will initially mount the slave NFS drive properly upon script execution, that non-breaking space prevents slave NFS drives from re-mounting properly after a slave server reboots, at least on Ubuntu 16.04. Running ` sed -i.bak 's/\xC2\xA0/ /g' slave_nfs_node.sh` on the script prior to execution will fix that up quickly.
The script in the repo just needs to be patched with a regular space.
$ tail -n 1 slave_nfs_node.sh | hexdump -C
00000000 65 63 68 6f 20 22 24 4d 41 53 54 45 52 5f 49 50 |echo "$MASTER_IP|
00000010 3a 2f 64 6f 63 6b 65 72 64 61 74 61 2d 6e 66 73 |:/dockerdata-nfs|
00000020 c2 a0 2f 64 6f 63 6b 65 72 64 61 74 61 2d 6e 66 |../dockerdata-nf|
<snip>