Nickolai Zeldovich <[EMAIL PROTECTED]> writes: > On 29 Jul 2002 19:04:22 +0200, Jimmy Engelbrecht <[EMAIL PROTECTED]> wrote: > > > > Empty CellServDB entries are not allowed by OpenAFS (even though they > > > are allowed by arla), thus resulting in the odd behavior you're seeing. > > > Why do you want to have empty CellServDB entries > > > > I want them in /afs even if they have not accessed yet. > > That's fine. They will be in /afs because they are mountpoints in > root.afs. When you try to access them, the cache manager will try > to evaluate the mountpoint, causing an AFSDB lookup for the cell in > question. So, in this situation you don't need empty CellServDB > entries. > > > So if i use want to use AFSDB-records i have to run dynroot ? > > I see no reason for that dependency. > > No, there's no dependency; I was similarly confused by your mention > of dynroot in your message. The two features are fairly independent. > > If you are using _both_ dynroot and AFSDB, there is a reason for empty > CellServDB entries: to tell dynroot to make mountpoints for those cells > before you try to access them, but fetch the volume server information > from AFSDB records. (Same reason as you mention above for non-dynroot > case, except I believe it only applies here and not there.)
You are right and i am wrong. A empty CellServDB will do what i want and make me happy. /Jimmy _______________________________________________ OpenAFS-devel mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-devel
