State of files.pharo.org
Locked
 
|
I'm trying to download images from files.pharo.org. Either it takes a lot of time or get something like
Downloading the latest 70 Image: error: invalid compressed data to inflate image.7vLfZK/Pharo7.0-32bit-6bff78c.image image.7vLfZK/Pharo7.0-32bit-6bff78c.sources bad CRC 724d27ab (should be f4d1baa8) That is really sad. What use is this wonderful environment if nobody can download it? We need to sort that out. So is this hosted in Inria data center? Norbert |
|
|
Maybe this is a transient problem, because we had some problems this morning to dl files from files.pharo.org from Cameroun.
Envoyé de mon iPhone
|
|
|
In reply to this post by NorbertHartl
Yes, and the admins can not find any problem. They closed the ticket I opened, but there is a new ticket open since last week… not a good situation. Marcus |
|
|
Yes, it is hosted at Inria datacenter. They are investigating the problem. I have an audio conf with them tomorrow on this pb. I will give a feedback after this meeting. Christophe |
Re: State of files.pharo.org
|
|
Thanks Christophe because Peter reported the same when I visiting him
and this is slowly killing us. On Thu, Oct 5, 2017 at 2:18 PM, Christophe Demarey <[hidden email]> wrote: > > Le 5 oct. 2017 à 13:50, Marcus Denker <[hidden email]> a écrit : > > > > On 5 Oct 2017, at 13:35, Norbert Hartl <[hidden email]> wrote: > > I'm trying to download images from files.pharo.org. Either it takes a lot of > time or get something like > > Downloading the latest 70 Image: > http://files.pharo.org/get-files/70/pharo.zip > error: invalid compressed data to inflate > image.7vLfZK/Pharo7.0-32bit-6bff78c.image > image.7vLfZK/Pharo7.0-32bit-6bff78c.sources bad CRC 724d27ab (should be > f4d1baa8) > > That is really sad. What use is this wonderful environment if nobody can > download it? We need to sort that out. > > So is this hosted in Inria data center? > > > Yes, and the admins can not find any problem. They closed the ticket I > opened, but there is a new > ticket open since last week… not a good situation. > > > Yes, it is hosted at Inria datacenter. > They are investigating the problem. I have an audio conf with them tomorrow > on this pb. I will give a feedback after this meeting. > > Christophe > > |
|
|
We have to circumvent the DevOp mental model :) From their point of view it goes like this
"I cannot find a reason why it shouldn't work….so it should. Ticket closed!" Just ignoring the little effect THAT.IT.DOES.NOT.WORK Same story everywhere :) Norbert > Am 06.10.2017 um 10:01 schrieb Stephane Ducasse <[hidden email]>: > > Thanks Christophe because Peter reported the same when I visiting him > and this is slowly killing us. > > On Thu, Oct 5, 2017 at 2:18 PM, Christophe Demarey > <[hidden email]> wrote: >> >> Le 5 oct. 2017 à 13:50, Marcus Denker <[hidden email]> a écrit : >> >> >> >> On 5 Oct 2017, at 13:35, Norbert Hartl <[hidden email]> wrote: >> >> I'm trying to download images from files.pharo.org. Either it takes a lot of >> time or get something like >> >> Downloading the latest 70 Image: >> http://files.pharo.org/get-files/70/pharo.zip >> error: invalid compressed data to inflate >> image.7vLfZK/Pharo7.0-32bit-6bff78c.image >> image.7vLfZK/Pharo7.0-32bit-6bff78c.sources bad CRC 724d27ab (should be >> f4d1baa8) >> >> That is really sad. What use is this wonderful environment if nobody can >> download it? We need to sort that out. >> >> So is this hosted in Inria data center? >> >> >> Yes, and the admins can not find any problem. They closed the ticket I >> opened, but there is a new >> ticket open since last week… not a good situation. >> >> >> Yes, it is hosted at Inria datacenter. >> They are investigating the problem. I have an audio conf with them tomorrow >> on this pb. I will give a feedback after this meeting. >> >> Christophe >> >> |
|
|
Often the problem is not that they cannot find a reason why, but before that, they cannot even see it, i.e. they can't put the hands around it. Its human nature that we deal with the world through our own experiences. If a problem can't be seen, how can it be solved.
Without know what actions they may have tries, I'd generally say it seems that it can't be solved with local testing since that is a different environment. Ideally your DevOps guys could get on a machine in a remote corner of the world exhibiting the problem, to help them "see" the problem. That said, right now for me from different location its not too bad. Perhaps it only get in a twist with lagged/dropped packets on poor networks. Do they test using tools to simulate poor networks? cheers -ben On Fri, Oct 6, 2017 at 4:12 PM, Norbert Hartl <[hidden email]> wrote: We have to circumvent the DevOp mental model :) From their point of view it goes like this |
|
|
In reply to this post by demarey
Hello,
I just finished the meeting with a guy from Inria datacenter. From Inria, we do not see any problem with files.pharo.org. We do not have tools ready to simulate poor networks nor remote network sensors available :( That’s why if some of you could help to collect some data to guess what is the problem, it would be great: If you can run the commands listed there: https://pad.inria.fr/p/IvdGLh650VN8lkoN and also put the results at the bottom of the page, it would be very helpful. In the same time, I created a project on github to run these command on travis CI: https://travis-ci.org/demarey/test-files-pharo-org. With all this data, I hope they can spot the problem quickly. Thanks, Christophe |
|
|
I've added my testing info in the shared notepad. The download of Pharo6.0-linux.zip took 11m 33s on a 30Mbps connection (I've looked while it was downloading and for the first half it was downloading at 40/50 Kbs), so I think it's representative of the slowness we get. As I previously said it's not always slow, sometimes I works fine. Regards, Gabriel On Fri, Oct 6, 2017 at 11:07 AM, Christophe Demarey <[hidden email]> wrote: Hello, |
|
|
Thanks a lot for the data.
Christophe
|
Re: State of files.pharo.org
|
|
In reply to this post by Ben Coman
On 06-10-17 16:00, Ben Coman wrote:
> Without know what actions they may have tries, I'd generally say it seems > that it can't be solved with local testing since that is a different > environment. Ideally your DevOps guys could get on a machine in a remote > corner of the world exhibiting the problem, to help them "see" the > problem. That could be easily and cheaply done by using some of the vps providers that let you work with a machine in a datacenter of your choosing. Anyway, what might help the Inria team is a traceroute from the slow connection Stephan |
«
Return to Pharo Smalltalk Developers
|
1 view|%1 views
| Free forum by Nabble | Edit this page |