Re: Is anyone at perlmonks aware that...
by ikegami (Patriarch) on May 19, 2011 at 06:53 UTC
|
>nslookup perlmonks.org
Server: home
Address: 10.0.0.1
Non-authoritative answer:
Name: perlmonks.org
Addresses: 209.197.123.153
66.39.54.27
216.92.34.251
209.197.123.153 and 66.39.54.27 work fine, but with 216.92.34.251:
>perl -MIO::Socket::INET -e"my $s = IO::Socket::INET->new(PeerAddr=>sh
+ift, PeerPort=>80); $s->print(<>); print while <$s>;" 216.92.34.251
GET / HTTP/1.1
Host: perlmonks.org
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.1
+7) Gecko/20110420 Firefox/3.6.17
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.
+8
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 115
DNT: 1
Connection: keep-alive
Cache-Control: max-age=0
^Z
HTTP/1.1 200 OK
Date: Thu, 19 May 2011 06:48:47 GMT
Server: Apache/2.2.14
Content-Length: 0
Keep-Alive: timeout=5, max=100
Connection: Keep-Alive
Content-Type: text/plain
The problem occurs for Host: perlmonks.org and www.perlmonks.org (the only two I tried), but "/" is the only URL I found for which the problem manifests itself.
It's not a network issue. The server is reached and gives a reply. The trace for every IP is the same save for the terminating IP.
| [reply] [d/l] [select] |
|
| [reply] |
Re: Is anyone at perlmonks aware that...
by BrowserUk (Patriarch) on May 19, 2011 at 03:03 UTC
|
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.
| [reply] [d/l] |
|
are you using safe browsing feature? Maybe that is having a timeout
| [reply] |
|
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.
| [reply] |
|
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.
| [reply] |
|
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.
| [reply] [d/l] |
|
|
|
|
Re: Is anyone at perlmonks aware that...
by roboticus (Chancellor) on May 19, 2011 at 02:13 UTC
|
7stud:
I've been having similar issues. I've been switching to perlmonks.com when I encounter that. I thought it was my company's network, though.
...roboticus
When your only tool is a hammer, all problems look like your thumb.
| [reply] |
Re: Is anyone at perlmonks aware that...
by jffry (Hermit) on May 19, 2011 at 17:45 UTC
|
ikegami's output shows Content-Length: 0, so the web server is perfectly aware it is serving nothing and seems happy with that result (thus 200 OK HTTP status). So what program is generating the content for the / URI? Does it log anywhere?
This also might not be happening on all servers in the cluster. Notice that BrowserUK's output shows a different version of Apache, and, more importantly, the default Content-Type is text/html but on ikegami's it is text/plain (even though it is 0 bytes). This is another hint that the program which generates the content for / URI is having trouble because it isn't changing the Content-Type away from Apache's (probably) default of text/plain.
| [reply] |
Re: Is anyone at perlmonks aware that...
by ikegami (Patriarch) on May 19, 2011 at 02:48 UTC
|
It happened to me about 8 times today. Reloading the page always worked. It's like the server never received the request because I had to repeat POST requests. It only affect PerlMonks, but I don't think it's specific to a give URL, node or domain. I don't know if it affects one or more servers. | [reply] |
Re: Is anyone at perlmonks aware that...
by CountZero (Bishop) on May 19, 2011 at 06:18 UTC
|
| [reply] [d/l] [select] |
Re: Is anyone at perlmonks aware that...
by Anonymous Monk on May 19, 2011 at 02:12 UTC
|
| [reply] |
Re: Is anyone at perlmonks aware that...
by choroba (Cardinal) on Jun 01, 2011 at 14:12 UTC
|
I usually see the blank page as well. Will try to remember to check some details of the page next time.Update: $ wget -SO- http://www.perlmonks.org/
--2011-06-01 23:55:14-- http://www.perlmonks.org/
Resolving www.perlmonks.org... 216.92.34.251, 66.39.54.27, 209.197.123
+.153
Connecting to www.perlmonks.org|216.92.34.251|:80... connected.
HTTP request sent, awaiting response...
HTTP/1.1 200 OK
Date: Wed, 01 Jun 2011 21:55:14 GMT
Server: Apache/2.2.14
Content-Length: 0
Keep-Alive: timeout=5, max=100
Connection: Keep-Alive
Content-Type: text/plain
Length: 0 [text/plain]
Saving to: `STDOUT'
[ <=> ] 0 --.-K/s in 0s
2011-06-01 23:55:14 (0.00 B/s) - `-' saved [0/0]
| [reply] [d/l] |
|
Bypassing DNS round-robin in my wgets...
In the last few minutes, when I get content from 216.92.34.251, 100% of the time it is zero bytes (bad).
Also, at this time, when I get content from 209.197.123.153 and 66.39.54.27, the content is a typical 68K page 100% of the time. (good)
Is there an admin who can look at md5sums or diffs of the programs which generate content at the bad server vs the good servers?
UPDATE: I am only referring to a GET of the "/" URI.
| [reply] |
|
I don't think an md5sum between mod_perl 1 and mod_perl2 would be very informative. :)
Thanks for reporting the IP addresses involved. Too often, people report something like "perlmonks.org is broken but perlmonks.com works" which is much less helpful (since I can't query their system to see which of the three IPs associated with perlmonks.org is currently being used by their browser nor which of the exact same 3 IPs associated with perlmonks.com is currently being used by their browser).
The zero-byte results that I have been able to reproduce (for certain nodes when viewed by certain people but not on the problematic 216 IP) have been due to core dumps. Meanwhile, the logs on the 216 IP have been completely unhelpful. I'm tempted to just request that the 216 be removed from the DNS for now, though I'm hesitant because it could lead to more over-all site slowness.
So far, it seems that the problem shows up for a fairly short while and then disappears. That makes me think it might be worth requesting an Apache config change so mod_perl children don't live quite so long.
I'm pretty sure pair.com monitors the web services (and I suspect the blank pages might trigger pair.com to restart Apache which could explain why the problem goes away fairly quickly). I should try to find if that monitoring data is actually available to me and might contain any useful information in diagnosing the problem.
| [reply] |
Re: Is anyone at perlmonks aware that...
by Aim9b (Monk) on May 19, 2011 at 17:52 UTC
|
When I get the blank screen, the view source shows me this...
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">>
<HTML><HEAD>>
<META content="text/html; charset=windows-1252" http-equiv=Content-Type></HEAD>>
<BODY> </BODY></HTML>
| [reply] |
|
I suspect that every bit of that particular "view source" result is actually just in the imagination of the browser. The browser "view source" feature has long been a bit of a lie, generating something that is rather similar to the actual source downloaded but that is actually at least partially generated by "reverse parsing" the data structure the browser built from the actual HTML source (and probably javascript manipulations).
| [reply] |
Re: Is anyone at perlmonks aware that...
by luis.roca (Deacon) on May 19, 2011 at 21:13 UTC
|
| [reply] |
|
I don't believe that this is related to what most people are reporting. For one thing, most of the symptoms reported here actually indicate that a response was received (otherwise the browser behavior would be perhaps a time-out and then certainly a browser-supplied error page).
For another, none of the IPs you show are tied to www.perlmonks.org (nor www.perlmonks.com, since it uses exactly the same IP addresses, nor versions w/o the "www." prefix).
Only the last one even has a reverse DNS entry (at the moment, from here). It appears to be ads.perlmonks.org, a hostname that I know nothing about. It appears to host a bad version of the site (scary). (Update: Ah, it'd be good to improve that reverse IP entry. Sorry I missed matching that one IP.)
At a guess, did you get these IPs by looking at what your browser said it was waiting for while it appeared to be trying to load this web site? I am guessing that those are IPs of advertisement servers.
Just FYI, if anybody wants to report network connectivity problems, then 'ping' results are fairly useless. traceroute (tracert on Windows) results can actually be somewhat informative.
| [reply] |
|
$ nslookup perlmonks.org
Server: 192.168.3.125
Address: 192.168.3.125#53
Non-authoritative answer:
Name: perlmonks.org
Address: 216.92.34.251
Name: perlmonks.org
Address: 66.39.54.27 <-------
Name: perlmonks.org
Address: 209.197.123.153
| [reply] [d/l] |
Re: Is anyone at perlmonks aware that...
by mje (Curate) on Jun 30, 2011 at 12:25 UTC
|
I have been getting what seems an empty page for the last few days. Switching to perlmonks.com solved the problem for me and until I saw this thread I just assumed something was wrong on perlmonks.org. Hitting refresh or even ctrl/refresh in Chromium still gave me an empty page. With live HTTP headers in firefox I see:
http://www.perlmonks.org/
GET / HTTP/1.1
Host: www.perlmonks.org
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-GB; rv:1.9.2.18) Gecko
+/20110628 Ubuntu/10.10 (maverick) Firefox/3.6.18 GTB7.1
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.
+8
Accept-Language: en-gb,en;q=0.8,de-de;q=0.6,fr;q=0.4,fr-fr;q=0.2
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 115
Proxy-Connection: keep-alive
If-Modified-Since: Thu, 30 Jun 2011 12:21:10 GMT
If-None-Match: "129595-ebb2-4a6ecefa67980"
HTTP/1.0 200 OK
Date: Thu, 30 Jun 2011 12:23:12 GMT
Server: Apache/2.2.14
Content-Length: 0
Content-Type: text/plain
X-Cache: MISS from andvari.easysoft.local
X-Cache-Lookup: HIT from andvari.easysoft.local:8080
Via: 1.1 andvari.easysoft.local:8080 (squid/2.7.STABLE7)
Connection: keep-alive
Proxy-Connection: keep-alive
Content-Length 0 | [reply] [d/l] |
Re: Is anyone at perlmonks aware that...
by 7stud (Deacon) on May 19, 2011 at 21:21 UTC
|
| [reply] |
|
I am having the same problem today.
It seems that if I google it first it works....
| [reply] |
Re: Is anyone at perlmonks aware that...
by locked_user sundialsvc4 (Abbot) on Jun 01, 2011 at 12:47 UTC
|
This definitely is a problem, and furthermore it is quite specific to this site. It started happening about a month ago. Nobody’s “seeing things” here.
I find that resetting my local router fixes the issue, so to speak, and that otherwise it might go on for days, no matter what browser(s) I might be using and no matter if I restart the computer. Other computers in the same network are not affected. If that helps ...
| |
|
| [reply] |
Re: Is anyone at perlmonks aware that...
by locked_user sundialsvc4 (Abbot) on Jun 30, 2011 at 11:58 UTC
|
Very interesting . . .
I got the “white screen o’ death” again this morning, and decided (since I could not reach perlmonks) to “Google it.” Which, in turn, led me to this page ... http://www.perlmonks.org/index.pl?node_id=905600. (In other words, at PerlMonks.)
The page popped right up!
I happened to have the “white screen” still open at the same time, and I reloaded it. White screen again. Reloaded “this page” again ... it popped right up. White here, Content there, both at the same time.
So, I deduce that there must be some kind of issue with regard to server-side caching or a proxy-server. HTH ...
I have already noticed that, when the white-screen does happen, it always happens “instantly.” There is no delay at all, unlike Perlmonks itself which (despite its beefy servers) sometimes takes several seconds to respond. And it does make sense: the page most-likely to be cached, hence to cause problems if cached incorrectly, is the opening (“empty URL”) home screen.
| |
Re: Is anyone at perlmonks aware that...
by locked_user sundialsvc4 (Abbot) on Jul 04, 2011 at 12:44 UTC
|
I have observed the “blank page” behavior for several days now, enough to make the hypothesis that a work-around to the problem is to bookmark, e.g. http://www.perlmonks.org/index.pl, or to Google and follow the direct “The Monastery Gates” URL ... specifically asking for the content instead of expecting the site to figure out a default. A preponderance of evidence seems to be growing, in my mind at least, that this serves as a workaround ... and, hence, perhaps a pointer to the underlying root cause.
I am growing more certain that this is a server-side cache related problem that appears to be specific to empty URLs.
HTH ...
| |