On Fri, Sep 21, 2012 at 12:47 PM, Troy Benjegerdes <ho...@hozed.org> wrote:
> I forgot part of my excercise:
>
> run the tests on an AFS filesystem from a machine that is
> running (production) AFS server instances. The runtests -o
> output shows a couple of IP addreses from my servers.

Well, since right now the servers use a fixed port, if the fixed port
is busy, the tests fail.

That can't be fixed until the service ports are discovered instead of fixed.

So any failures there can be fixed trivially: don't run them on a
machine already running AFS servers.

> I see the auth/superuser failure on a number of machines
> that work just fine as AFS clients.
>
> On Thu, Sep 20, 2012 at 11:43:14AM +0100, Simon Wilkinson wrote:
>>
>> On 20 Sep 2012, at 07:42, Troy Benjegerdes wrote:
>>
>> > So here's a fun excercise:
>> >
>> > build the latest master on a local filesystem, run make check, get
>> > something like this;
>>
>> [ snip ]
>>
>> If you run each of the failing tests with runtests -o (having set 
>> LD_LIBRARY_PATH appropriately), you'll get a more complete set of output, 
>> which may provide clues as to why these tests are failing. It looks like 
>> your vos failure is because your machine's network stack is setup with the 
>> machine's hostname resolving to a loopback address. As this isn't a valid 
>> configuration for a ubik server, any ubik test (which includes the 
>> volser/vos test) will fail.
>>
>> I don't have any immediate ideas about the auth/superuser failure, but it 
>> looks like that's some kind of connectivity issue as well, as the failing 
>> tests are those which try an RPC from a client to a server using superuser 
>> authentication.
>>
>> Cheers,
>>
>> Simon
>>
>> _______________________________________________
>> OpenAFS-devel mailing list
>> OpenAFS-devel@openafs.org
>> https://lists.openafs.org/mailman/listinfo/openafs-devel
> _______________________________________________
> OpenAFS-devel mailing list
> OpenAFS-devel@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-devel
>



-- 
Derrick
_______________________________________________
OpenAFS-devel mailing list
OpenAFS-devel@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-devel

Reply via email to