Closed dr0i closed 6 years ago
See also hbz/lobid-resources#621 which will update lobid-resources to es-5.6.3 on weywot5.
We had another nfs-caused incident as reported in hbz/lobid-resources#586. Since the migration of lobid to the new weywot-cluster is done we are ready to get rid of nfs. Will configure all servers to use local harddisks instead of the error-prone nfs.
Done. Most server will use a NFS-mount anyway, but nomore the mount with the big data. Think this mount was the bad one, because the other NFS-mounts are used on most other hbz server and they hadn't had that 7 h breakdown, don't they? Gaia, serving as a git-daemon, will stay with the bad NFS-data-mount because it uses that to store its repos.
As there is not much free disk space on our elasticsearch cluster
gaia-aither
there is a dependency on anfs-mount
for loading the hbzXmlClobs. On the other hand we have our weywot's-server which have much free hard disk and also enough RAM since some months now. So we could migrate to get rid of the nfs-mount.