Bug#624414: changes in squeeze affecting a smooth NFS upgrade/use

[Luk Claes <luk@debian.org>]

Package: release-notes
Severity: normal

Hi

I keep telling me I should write something that could be included into the Release Notes about things people should be aware of when using NFS in squeeze. Instead of trying to have a coherent text I'll just list some topics:

- default NFS uses version 4 in squeeze, one can work around failures to mount because of this by mentioning the -vers=3 option explicitly when mounting (or in fstab)
- when using Kerberos default encryption has changed, one can work around it by specifying an older encryption type explicitly (-sec=...) when mounting
- the chance that NFS related daemons are not started at boot are increased both for NFS clients as well as NFS servers, can be worked round by restarting portmap (or rpcbind), nfs-common and nfs-kernel-server in that order for the ones that are installed
- NFSv4 needs the idmap daemon to be started, can be worked around by updating /etc/default/nfs-common (and /etc/default/nfs-kernel-server) to have NEED_IDMAPD=yes and restart nfs-common (and nfs-kernel-server)
- for switching the server exports to NFSv4 one needs to specify a root of all exported filesystems specified by including the fsid=0 (or fsid=root) option, one does not need to specify fsid for any other exported filesystem unless it has no uuid (in that case use a small integer that will be used like the uuid)

Cheers

Luk