in reply to Is anyone at perlmonks aware that...

I'm also getting a blank screen for perlmonks.org in my browsers (Opera & Firefox), and attempts to show source fail completely.

However, if I use lwp-request, I get the following headers:

c:\test>lwp-request -Sed http://perlmonks.org/ GET http://perlmonks.org/ --> 200 OK Connection: close Date: Thu, 19 May 2011 02:58:07 GMT Accept-Ranges: bytes ETag: "129595-f7d9-4a3982a11aac0" Server: Apache/2.2.17 Content-Length: 63449 Content-Type: text/html Last-Modified: Thu, 19 May 2011 02:57:23 GMT Client-Date: Thu, 19 May 2011 02:58:06 GMT Client-Peer: 209.197.123.153:80 Client-Response-Num: 1 Link: </css/common.css>; rel="stylesheet"; type="text/css" Link: <?node_id=204962>; rel="stylesheet"; type="text/css" Link: </favicon.ico>; rel="icon" Title: PerlMonks - The Monastery Gates X-Meta-Description: A community committed to sharing Perl knowledge and coding tips. The site contains questions and answers, useful snippets, and a library of code. X-Meta-Keywords: perl, mod_perl, regular expressions, regexp, xp whoring, CGI, programming, learning, tutorials, questions, answers, examples, vroom, tim, node, experience, votes, code

And if I omit the -d flag, I get just over 64k of html that looks very like a normal Monastery Gates page.

Clever that. A non-browsable web page :)

Update: for me the above results are totally consistant on perlmonks.org and perlmonks.com from both browsers continually. Refreshing never results in any change.

However, using the IP address directly works fine in both browsers. At that point, I'm stumped unless it is something to do with DNS--but how, why?


Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.

Replies are listed 'Best First'.
Re^2: Is anyone at perlmonks aware that...
by Anonymous Monk on May 19, 2011 at 04:07 UTC
    are you using safe browsing feature? Maybe that is having a timeout
      are you using safe browsing feature?

      Nope. And if that were the cause, then why would using the IP work?

      And the request completes (status bar says 'done') in less than a second, so I don't think that it is timing out.

      One thing that is interesting is that both urls work perfectly in an ancient version of opera (6.06) with javascript disabled.


      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.
Re^2: Is anyone at perlmonks aware that... (browser vs lwp-request)
by tye (Sage) on Jun 02, 2011 at 18:36 UTC

    My guess is that your browser(s) was caching the one (intermittently) bad IP at that time (while lwp-request was using one of the other two IPs). If you had managed to run lwp-request against the bad IP during the time of its badness, I think you would have gotten zero bytes of content in response (perhaps not even HTTP headers).

    Apache can be configured such that getting "Host: 216.92.34.251" produces different results than getting "Host: perlmonks.org", but that isn't currently the case on any of the IPs currently assigned to the PerlMonks DNS names.

    - tye        

      Sorry, but that doesn;t seem to be the case.

      When I just read your post (accessed vis http://209.197.123.153/), I opened a new tab and typed "perlmonks.org" into Opera and again I got a blank page that competes immediately.

      I then typed:

      lwp-request http://perlmonks.org/

      And got the front page.

      I then returned to the tab I opened and hit F5 which in Opera forces a full refresh and got another blank page.

      Finally, I fired up Firefox and went intooptions/privacy/empty recent history. Then I tried perlmonks.org and got a blank page.

      This is not a spurious problem limited to me or just one browser, nor a transient one. It cannot be so easily dismissed.


      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.

        I think the best approach is to leave DNS completely out of the picture.

        Currently, http://216.92.2.139/~perl2 http://216.92.34.251/index.pl (which is also accessible via sending the appropriate Host: header and accessing the IP address directly) serves a blank page for me, most likely because the cached frontpage has been generated empty.

        You don't appear to have understood most of what I wrote as I wasn't dismissing the complaint and your reproduction steps had really nothing to do with what I was pointing out.

        My point: lwp-request is choosing to use a different IP than your browsers are using when given the same host name. (I'm a little surprised that your browsers are so insistent on using the same IP, but that may just be a Windows thing.)

        [ But the fact that only "/" is constantly failing (on that IP) was something that I only clued into after re-reading ikegami's reply (in part due to a large number of "oh, now it is working" replies here and via other media).

        (Update: Ah, it appears that there are two different problems just on that one IP. One is "/" (the Monastery Gates) being blank (always, for everybody) and the other is lots of pages being blank only every so often (at least for some people -- perhaps based on the user's IP address). And this second probably should now be fixed, thanks to Corion. ) ]

        - tye