[OpenAFS-devel] --fast-restart?

[Jeffrey Hutzelman]

On Tuesday, August 30, 2005 02:46:07 PM -0400 chas williams - CONTRACTOR 
<chas@cmf.nrl.navy.mil> wrote:

> the NEWS file says:
>
> ** If the --enable-fast-restart flag is given when configuring AFS,
>    then the salvager supports the -dontsalvage flag which causes it to
>    exit without salvaging any volumes.  If this is configured into the
>    third command of a fs process, then the fileserver will start without
>    salvaging.  It will fail to attach volumes that need salvaging and they
>    can be salvaged manually.  This provides significantly better server
>    startup performance at the cost of administrative complexity.
>
> but what i dont understand is how volumes get marked as needing
> salvaged?  salvager doesnt do it and fileserver doesnt do it.
> what does?
>
> if i kill -9 a fast-restart fileserver it reattaches all the volumes
> without a complaint.  this seems "bad".


So, the idea here is that you don't force a full salvage of every volume on 
a partition just because the fileserver has crashed.  If the fileserver 
detects a problem with a volume, it will still mark the volume "needs 
salvaged" and take it offline, and since that flag is persistent, that 
volume will _stay_ offline until it gets salvaged -- presumably by an admin 
running 'bos salvage'.

Note that the third command on an fs bnode is used only for automatic 
salvages triggered by the bosserver when the fileserver crashes or when it 
finds the flag file indicating that bnode was not cleanly shut down.  The 
'bos salvage' command works by creating a 'cron' bnode with frequency 
'once' and the command AFSDIR_CANONICAL_SERVER_SALVAGER_FILEPATH.

-- Jeff