[RADIATOR] configuration validation

Eric W. Bates ericx at whoi.edu
Tue Mar 14 13:03:58 UTC 2017


Thank you.

I have to apologize: something in the mail chain blocked several 
messages; so I did not see your quick replies.

On 03/14/2017 05:00, Heikki Vatiainen wrote:
> On 13.3.2017 16.54, A.L.M.Buxey at lboro.ac.uk wrote:
>
>>> There's -c option for radiusd. It will read the configuration and
>>> checks what it can. It does not, for example, attempt to connect to
>>> SQL databases, so it will not catch those types of problems.
>>
>> it does appear to do RADSEC connections though. maybe SQL connectivity
>> should
>> be part of the test (or an option with another flag?)
>
> Correct, it seems that RADSEC hosts are activated and try to connect.
> Good catch, I'll see if this should be changed for config check.
>
> A flag for testing external connections is probably needed since, for
> example, what we do in some cases is to run tests in a separate
> environment where connection attempts would not work. But I can see that
> in some cases the connectivity testing would be a better option.
>
> We'll need to think about this.
>
>>> There's actually work ongoing to enhance configuration checking.
>>> Currently the output from -c is more for human consumption, so the
>>> exit code, for example, does not reflect the check results. We are
>>> looking at returning different exit codes depending of level of
>>> problem (warning, error) from -c run.
>>
>> yes, interesting timing on this query since I've been recently in contact
>> with OSC over this feature :)
>
> Indeed :) What we have used is the wrapper that checks logs, but exit
> codes would be a useful addition too.
>
>>> What you could do now is to wrap radiusd -c invocation with a script
>>> that greps errors and warnings and then returns non-zero exit code.
>>
>> ...if you get output ;-)
>
> To clarify this a bit for the list members, a couple of cases, such as
> badly formatted <Handler ...> selector is not logged during -c run yet.
> So this is under work too.
>
> Thanks,
> Heikki
>

-- 
Clark 159a, MS 46
508/289-3112

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4188 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.open.com.au/pipermail/radiator/attachments/20170314/6f67a7bf/attachment.p7s>


More information about the radiator mailing list