On Sun, May 17, 2009 at 12:18 AM, Atro Tossavainen
<[email protected]> wrote:
> Derrick Brashear wrote:
>
>> CellAliases file is your friend, and ameliorates this argument.
>>
>> I'd guess you're actually content with dynroot, and don't know it yet. ;)
>
> I'm getting the impression that doing that would be replacing a bit of
> configuration that's currently global and in AFS itself with something
> that is per-client and in a file on each client.

CellAliases would be a distributed file, just as CellServDB is, but,
you have a point.

> I'm not sure yet
> this is a good idea.  On the other hand, I do acknowledge the choice
> to attach the client to a particular cell rather than dynroot is also
> a configuration item in a local file on each client.

That too.

> How does this affect authenticating logins against AFS?  (Yes, spare
> me the argument against kaserver.  I *know*.  We might not be using
> AFS by the end of this year anyway, The Them want to replace it with
> SMB/CIFS/NFSv4/whatever.)

It doesn't affect it positively, negatively, or sidelong.

Derrick
_______________________________________________
OpenAFS-devel mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-devel

Reply via email to