[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Issue 13.7: copy-config
At 04:43 PM 3/29/2004, Rob Enns wrote:
>As I recall the main reasons copy-config was invented are:
>
>1) copy <running> to <startup>
>
> ie. if (#startup)
> copy <running> to <startup>
>
>2) copy named configs around
>
>I agree that the other cases you enumerated are pretty useless.
>
>Since NETCONF won't have named configs, that leaves only
>copy <running> to <startup>. If there's a better way to do
>that, or if it's acceptible to do it by editing the startup
>directly, let's dispense with copy-config.
Originally we were using <commit> for both candidate --> running
and running --> startup. I think we changed this to save 2
parameter values (source, target). Phil argued that the copy
from candidate --> running can fail, but the copy from
running --> startup cannot fail (actually that's not true,
it's just not likely).
I don't know of any boxes that have both #candidate and
#writable-running capabilities, but I guess we have to
allow for that. I would suggest putting the 'source'
parameter back (the target is hardwired to running
or startup, depending on the source), and using <commit>
instead of <copy-config>. I think <commit> is more
clear and it's better if <commit> is in the base operation set.
Confirmed-commit can still be a #candidate extension.
>Rob
Andy
>> -----Original Message-----
>> From: owner-netconf@ops.ietf.org
>> [mailto:owner-netconf@ops.ietf.org] On Behalf Of Andy Bierman
>> Sent: Saturday, March 27, 2004 7:11 AM
>> To: netconf@ops.ietf.org
>> Subject: Issue 13.7: copy-config
>>
>> Hi,
>>
>> There are several problems with the copy-config command, as
>> it is currently defined:
>>
>> 13.7.2: too many options; no option discovery
>>
>> Most of the copy modes are optional:
>> - <running> as target
>> - <startup> as target
>> - remote to remote copy
>>
>> There is no way to tell what the agent actually supports
>> for source and destination parameters.
>>
>> 13.7.3: limited usefulness without user named databases and files
>>
>> Apparently, the only mandatory behavior an application can
>> count on is:
>> if (#candidate)
>> (1) copy <running> to <candidate>
>> if (#candidate && #url)
>> (2) copy remote-file to <candidate>
>> (3) copy <candidate> to remote-file
>> if (#candidate && #startup)
>> (4) copy <startup> to <candidate>
>> if (#url)
>> (5) copy <running> to remote-file
>> if (#startup && #url)
>> (6) copy <startup> to remote-file
>>
>> Mode (1) and (2) are redundant because they are available
>> via edit-config. It's not clear why anyone would want
>> to use mode (3). Mode (4) isn't that interesting since
>> not many products have #candidate and #startup capability.
>> Mode (5) and (6) are widely used by operators.
>>
>> 13.7.4: implicit format conversion not interoperable
>>
>> The format conversion "feature" of copy-config is not
>> well defined and application developers don't have any
>> way of knowing how XML <--> text translation will be
>> done on the agent. The mapping may be incomplete and
>> there is no way to tell which text or XML bits do not have
>> an appropriate mapping.
>>
>> 13.7.5: <url> mode not documented
>>
>> The text does not mention that <url> may be used in place
>> of one of the config targets. This just shows up in the
>> example.
>>
>> 13.7: Proposal
>>
>> Since this operation has several problems, and is not
>> very useful without user named configurations, it should
>> be removed and deferred until that feature is added in the
>> future.
>>
>>
>>
>>
>>
>> --
>> to unsubscribe send a message to netconf-request@ops.ietf.org
>> with the word 'unsubscribe' in a single line as the message text body.
>> archive: <http://ops.ietf.org/lists/netconf/>
>>
>>
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>