[db-wg] problems observed for - Re: Release 1.88 deployed to RC

Dear Ruediger, WG,

Thank you for helping to test the new release!

We understand your reasoning for a systematic and predictable way to refresh the RC Data. We will do a fresh (dummified) data sync with every major release to the RC environment in future. This requirement was not abundantly clear to us earlier, but we certainly have nothing against this if it helps you and others test better.

With regards to this release: we are currently syncing the latest (dummified) data to the RC environment - the process for this is a bit time consuming but we expect that the data will be there by the end of the day. Because we want to ensure that there is enough time to re-test this release, and we want to avoid changes just before the RIPE meeting if it can be helped, we plan to postpone deploying the current release to production to Monday 31 October.

Kind regards


Tim Bruijnzeels

Assistant Manager Software Engineering
RIPE NCC Database Group



> On 07 Oct 2016, at 18:53, Ruediger Volk, Deutsche Telekom Technik - FMED-41.. <rv at NIC.DTAG.DE> wrote:
> 
> Hello Alex,
>> Hello Ruediger,
>> 
>> It appears we just crossed paths. I sent you a reply to the ticket you
>> opened earlier.
> and I responded with more details before seeing your message on the list.
> 
>> As explained in my other message, the data that is available in the
>> Release Candidate is refreshed periodically, but certainly not with
>> every release.
>> The data that is currently in RC is at east several months old, which
>> could explain the differences you are seeing.
> For one the examples of differences that I reported seem not to be
> explained by this.
> 
> If the RC data is not refreshed in a documented systematic fashion
> we would need documentation of the last sync date for the RC data set;
> as is this is not given the straight forward assumption seems to be
> last sync was done at the time of the setup of the current RC.
> I just had a look at the RCEnvironment web page, and found text that 
> seems to explain when production data is synced into the RC:
> 
> "Even though the Release Candidate database contains a snapshot of the
> live data from the start of the testing period, ..."
> Seems "start of the testing period" can be read as "installation of
> the current RC" [or "publication of RC is ready for testing"].
> Looks to me like a promise of fresh data with each new RC -
> I probably did read this long ago and was happy about a reasonable
> choice and that it accomodates nicely the test scenarios that
> our applications offer.
> I don't remember any explicit discussion about this.
> 
>> We will also make sure that we update the page with the expected release
>> date of the Release Candidate as soon as we can.
>> 
>> My sincere apologies for any inconvenience caused.
> thanks for your responses, though I consider case and issues
> not closed and needing detail attention by appropriate staff.
> 
> Kind regards,
>  Ruediger
> 
> 
> Ruediger Volk
> 
> Deutsche Telekom AG -- Internet Backbone Engineering
> 
> E-Mail:  rv at NIC.DTAG.DE
> 
> 
>> Kind regards,
>> 
>> Alex Band
>> 
>>> On 7 Oct 2016, at 12:59, Ruediger Volk, Deutsche Telekom Technik - =
>> FMED-41.. <rv at NIC.DTAG.DE> wrote:
>>> =20
>>> Hi Tim, dear colleagues,
>>> =20
>>>> Dear working group,
>>>> =20
>>>> We have released whois version 1.88 to the Release Candidate (RC)
>>>> environment. We plan to deploy this version to production on Monday =
>> 10
>>>> October.
>>> =20
>>> did anyone else ALSO observe and report problems from testing with the=20
> =
>> 
>>> release candidate?
>>> We reported problems and added some related questions.
>>> The ticket number is NCC#2016100974 and after 24 hours we have not=20
>>> received any response and the ticket system does not document any NCC
>>> activety.
>>> =20
>>> With no response on reported problems I have to conclude that =
>> unresolved
>>> issues with 1.88 are known and deployment has to be postponed.
>>> =20
>>> We would have preferred to do the testing earlier - but as far as I =
>> know
>>> there has been no advance information for 1.88 scheduling.
>>> The release note page was showing TBD in all fields for 1.88
>>> days after RC deployment at least until Monday Oct 3rd (actually
>>> seems the update only happened Tuesday - potentially due to a comment
>>> I posted over the weekend)
>>> I previously pointed out that advance notice of rough schedule=20
>>> should be given in order to allow interested user parties to =
>> anticipate
>>> related activeties (such as testing) in their own (potentially tight)
>>> schedules - I think that argument contributed to including a field
>>> for planned RC deployment date on the release note/overview page.
>>> With surprise notice of RC availability on Thursday afternoon
>>> during turn of the month business activety peak days (and for our
>>> team a long weekend extending to Monday) did not offer a serious =
>> chance
>>> for early testing.
>>> =20
>>> Looking forward to responses on the reported problems and questions
>>> (i.e actual handling of the open ticket), and public announcement
>>> of rescheduled deployment taking into account resolution of open
>>> issues and sufficient time to assess problems/resolution.
>>> =20
>>> Best regards,
>>> Ruediger
>>> =20
>>> =20
>>> Ruediger Volk
>>> =20
>>> Deutsche Telekom AG -- Internet Backbone Engineering
>>> =20
>>> E-Mail:  rv at NIC.DTAG.DE