dev@glassfish.java.net

Re: major degradation in download speed for 'maven bootstrap-all'

From: Dinesh Patil <Dinesh.Patil_at_Sun.COM>
Date: Thu, 25 Jan 2007 14:19:53 -0800

Yes, Looks like downloading from download.java.net has become little
slower overall..

Here are some of the facts:

    * On Trunk (GlassFish V2): bootstrap-all goal takes *27 mins* on my
      Solaris_X86 machine.
    * On GlassFish V1 branch (SJSAS90_FCS_BRANCH): this goal still takes
      *25 mins* on same machine.
    * Difference on V2 is we have added glassfish-maven-repository URL
      with few additional binary components, but above difference is not
      that much.
    * I tried changing the order of the repositories, by adding first
      URL which most of the developers concerned, and able to get
      bootstrap-all goal duration to *17 mins*.
          o maven.repo.remote=http://download.java.net/javaee5/external/shared,http://download.java.net/javaee5/trunk/n
            ightly/${glassfish.os.name}/,http://download.java.net/javaee5/external/${glassfish.os.name}/,https://maven-
            repository.dev.java.net/nonav/repository,http://www.ibiblio.org/maven
    * I will be removing glassfishwiki.org as its not of any use because
      of unreliable connection.

if you have any suggestions to improve this further, please let me know.
thanks
Dinesh

Lloyd L Chambers wrote:

> Thank you Dinesh.
>
> Byron Nevins says he is seeing the problem too. The oddest part is
> the long pauses before a download begins; it might be that the
> throughput is reasonable, but there are big "naps" in between files.
>
> Lloyd
>
> On Jan 25, 2007, at 11:09 AM, Dinesh Patil wrote:
>
>> Yes, I am looking into this, will report back what's causing this
>> much delay in downloading the binaries from download.java.net..
>> thanks
>> Dinesh
>>
>> Lloyd L Chambers wrote:
>>
>>> All,
>>>
>>> This problem still persists today, Thursday, January 25. That
>>> makes it nearly a week.
>>>
>>> This morning it took 29 minutes (half an hour) to do 'maven
>>> bootstrap- all', with a failure at the end.
>>>
>>> Something has changed, and it's NOT my end. All other web sites
>>> are fast. Another developer has confirmed a problem.
>>>
>>> Lloyd Chambers
>>>
>>> On Jan 22, 2007, at 3:49 PM, Lloyd L Chambers wrote:
>>>
>>>> Speed of glassfish.dev.java.net for 'maven bootstrap-all' has
>>>> degraded massively, starting late last week. This is not over
>>>> VPN; it's direct access.
>>>>
>>>> Has a change been made to the script or server that would account
>>>> for this?
>>>>
>>>> I'm seeing long download pauses between jar files of 30 seconds
>>>> or more before each jar file starts to download ("Attempting to
>>>> download foo.jar"), then a slow download speed once the file starts.
>>>>
>>>> It used to run at 1 mega BYTE per second download, with very
>>>> short pauses between files. Now it often runs at something
>>>> around 30K bytes /sec, though the larger files sporadically up
>>>> their rate briefly to 1MB/sec.
>>>>
>>>> The net result is that 'maven bootstrap-all' can now take 10-15
>>>> minutes, instead of a minute or two.
>>>>
>>>> It's not my ISP; other sites remain very fast.
>>>>
>>>> Lloyd Chambers
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>