Re: Pb Megacam image transfert from CADC

From: Kastelic, William <William.Kastelic_at_nrc-cnrc.gc.ca>
Date: Sun, 25 May 2008 08:09:28 -0700

The router tests were completed at approx 6:15 PDT Saturday.

The problem Friday was due to a CANARIE router issue in Calgary, effectively causing all (most?) BC research/academic sites to disappear from the intl res nets.

The HIA router tests Saturday were to see if we could use at least our commodity link to reroute research traffic if a similar problem occurs in the future.

Will
Will

----- Original Message -----
From: JJ Kavelaars <jjk_at_hia.nrc.ca>
To: Data Oversite Group Group <dog_at_cfht.hawaii.edu>
Cc: Kastelic, William
Sent: Sun May 25 10:53:40 2008
Subject: Re: Pb Megacam image transfert from CADC

Hi Yannick,

To correct the problems we had with our router a few days ago, the
network group in Ottawa has been making some changes. I suggest that
we will not have a highly robust connection until late on Sunday. If
you can wail until then, that would likely be the least frustrating
for you, otherwise you may need to continue to re-try downloads until
they succeed.

JJ

On 25-May-08, at 6:19 AM, Yannick Mellier wrote:

>
> Dear John
>
> To keep you informed: I tried new transfers and got 3 more
> interruptions. Only 5 images transferred successfully over 12.
>
> Cheers
>
> yannick
>
> Ouellette, John wrote:
>> My apologies Yannick -- I just reread your message and realized
>> that the times you mentioned were exactly during the period of our
>> network problems. The download that never completed was likely
>> just not terminated correctly by curl. Since your subsequent
>> download using wget worked, curl should now work, too. I will
>> however still talk to our network admin about the problem.
>> John
>> ----- Original Message -----
>> From: mellier_at_iap.fr <mellier_at_iap.fr>
>> To: Ouellette, John
>> Cc: Yannick Mellier <mellier_at_iap.fr>; Data Oversite Group Group <dog_at_cfht.hawaii.edu
>> >
>> Sent: Sat May 24 13:22:57 2008
>> Subject: Re: Pb Megacam image transfert from CADC
>> Dear John
>> Thank you very much for your email and for looking at
>> the problem.
>> I started a transfer yesterday at about 8:00 pm, Paris local
>> time. About 10 files were transferred without problem. All are
>> transferred using "curl". At about 9:00 pm yesterday, I curl-ed
>> 967615p but it vever finished: about 2/3 were only transferred
>> this morning at 7:00 am, though it kept continuing to transfer
>> (no interruption over the night). In order to understand what
>> could happen I then went to CADC interface and got this image
>> interactively, using wget. It worked well
>> , but transferred the .gz image (I usually transfer the .fz image). I
>> interrupted the wget after about 10 mn this morning since I saw
>> thre was no problmen with wget and the .gz image. THis is
>> why you saw a "failure" after 5-10 minutes, I guess? In any case, I
>> stopped the curl also and gave up. I will try again tomorrow and
>> will
>> keep you informed on the transfer.
>> Cheers and thanks again,
>> yannick
>> >
>> > Hi Yannick -- There do not seem to be any problems on the CADC
>> end: a
>> > few tests show that data is still capable of being transferred
>> from our
>> > web servers at >60Mbytes/sec, and I can't see any failures on our
>> > storage nodes. According to our logs, there was one file
>> (967615p) that
>> > didn't transfer properly to you, but the failure occurred after
>> five or
>> > ten minutes, not five hours: there don't seem to be any problem
>> with
>> > transferring that file at this time.
>> >
>> > Could you send me a list of files that you're having problems with?
>> > Perhaps I'm not seeing those failures being logged.
>> >
>> > John Ouellette
>> >
>> > Yannick Mellier wrote:
>> >>
>> >> Dear JJ
>> >>
>> >> Is there any problem at CADC? It takes 5 hours to transfer
>> one
>> >> image.fz from CADC to Terapix.
>> >>
>> >> Cheers
>> >>
>> >> yannick
>> >>
>> >
>
>

Received on Sun May 25 2008 - 05:09:36 HST