…and this is a 
We're having some major performance problems, as is obvious from all the errors you're getting when you try to post comments. We had some substantial tweaks made to the code behind the scenes that was supposed to improve performance, but actually had exactly the opposite effect — now Scienceblogs is supposed to be bringing in some expert consultants very soon now to fix the problems, either rolling back the code or figuring out why we aren't getting any speed boost at all. I don't know when this will happen, since there are a lot of blogs here, and they all need to be patched somehow. We're all feeling a bit frustrated by the mess.
Almost always, though, when you post a comment, it is getting added to the database, even if it does bounce back with a timeout error or some other noise. Please don't just go back and repost the same comment again and again. Go back, reload the page, and see if your comment has appeared first.
Comments
Posted by: Celtic_Evolution | May 27, 2009 8:26 AM
Thanks for the update, PZ...
However, I won't even be the littlest bit surprised when we get a double-posted comment in this thread.
Posted by: Rorschach
|
May 27, 2009 8:31 AM
This dude offered some solution,doesnt seem to have worked I guess:
http://www.thinknew.ca/ssd/index.php/2009/05/25/scienceblogs-has-some-problems/
The fact that this has been going on for over a week now is rather embarrassing for the scienceborgs,methinks.
Posted by: Nerd of Redhead, OM
|
May 27, 2009 8:34 AM
Thanks for the comment PZ. Acknowledging the problem is always the first step to fixing it. My last double post was when I changed browsers (to Safari 4 beta) and forgot they had changed how they showed web activity.
Posted by: Ahcuah | May 27, 2009 8:40 AM
If you could pass this along to the powers-that-be (since email to the webmaster bounced!):
The "Last 24 Hrs" pages aren't updating, either.
Posted by: Benjamin Geiger | May 27, 2009 8:58 AM
Celtic_Evolution @ #1:
Come on! You know we'd never, ever, multiple post!
Posted by: Benjamin Geiger | May 27, 2009 9:00 AM
Celtic_Evolution @ #1:
Come on! You know we'd never, ever, multiple post!
Posted by: JDStackpole | May 27, 2009 9:01 AM
Maybe it is those guardian angels (see earlier post, 7:46 AM) wreaking vengeance for your cavalier attitude toward them, messing up the comments system just to show you who is in charge.
Posted by: Gruesome Rob | May 27, 2009 9:10 AM
Scienceblogs should make a change analogous to a feature of slashdot. There's a hidden identifier on submit forms that prevents submission of duplicates. If it sees the same identifier, the post is rejected.
Posted by: Mike K | May 27, 2009 9:12 AM
@7: Fool!
It is DOG, smiting the evil and twisted scienceblogs!
Do you not see HIS work when you see it?
man, will you be in trouble when HE comes...
Posted by: Pi Guy | May 27, 2009 9:16 AM
Damn. I was going to say the same.
Posted by: Epinephrine | May 27, 2009 9:27 AM
I blame DOS attacks by christwire.org! They obviously bear a grudge for having reduced their servers to smoldering wreckage.
Actually, since the trouble began before we invaded their servers they clearly had prescient knowledge of the events, and I'll have to break out Dr. Dan Streetmentioner's Time Traveler's Handbook of 1001 Tense Formations...
They wioll-haven-bearong a grudge forwhen presooning we mayan on-reducen their servers to smoldering wreckage.
Posted by: MPG | May 27, 2009 9:28 AM
˙˙˙ʎzzıp pǝǝɟ I
˙˙˙sɯǝlqoɹd ƃuıʌɐɥ sı ʎllɐǝɹ sƃolqǝɔuǝıɔS `ʍoʍ
Posted by: Amph
|
May 27, 2009 9:35 AM
The last few weeks I can't reach the blog at least 10% of the time. Seeing the title of this post, I expected that this would be the subject.
I wonder if am the only one having this problem?
Posted by: Zeno | May 27, 2009 9:36 AM
I commented several times in the last few days and got an error for every single one. Instead of re-posting, however, I simply opened another browser window and visited Pharyngula again. Every single time I saw that my comment had made it. And it helped me avoid double-posting.
As for Epinephrine's comment, "I'll have to break out Dr. Dan Streetmentioner's Time Traveler's Handbook of 1001 Tense Formations...", didn't you already do this in the future?
Posted by: Clemens | May 27, 2009 9:36 AM
No offense, but how many poster are there on ScienceBlogs and aren't there websites around with significantly more contributers but significantly less performance issues? :-P
Posted by: Sili
|
May 27, 2009 9:54 AM
bdɯ,
Who's Dizzy, and why are you feeding her?
Posted by: Rev. BigDumbChimp | May 27, 2009 9:54 AM
Curious as to how many people will read this and ignore it exactly like the error message?
Posted by: Epinephrine | May 27, 2009 9:59 AM
Don't be such a cynic :P
Posted by: Epinephrine | May 27, 2009 10:01 AM
Don't be such a cynic :P
Posted by: Naked Bunny with a Whip
|
May 27, 2009 10:08 AM
I've been doing a select-all-copy routine before hitting the Post button, though the server hasn't actually lost any of my little rabbit pellets yet. *waits for the cheers... and waits...*
Posted by: Jeff F | May 27, 2009 10:12 AM
Maybe it's time to give up on praying over the server.
It's odd that ScienceBlogs of all sites doesn't have any capable IT people.
Posted by: Rev. BigDumbChimp | May 27, 2009 10:13 AM
Epinephrine, I'm going to just assume that was on purpose and say nicely done.
Posted by: Ping "Chartebeat" Net | May 27, 2009 10:16 AM
It is MY evil doing!
Bwa! Hahaha!
I am also letting comments through (eventually, bwahaha) without an e-mail address!
For I am EVIL! BwaHa!
Posted by: Eric T | May 27, 2009 10:20 AM
Thought it was me or possibly an interruption of service attack by some other unappreciative group.
Posted by: Sven DiMilo | May 27, 2009 10:22 AM
It's a sad state of affairs indeed when parody double-posts are actually kind of funny (but only the first couple of times! Not funny from now on!).
Next parody apologies for parody double-posts?
Posted by: LT | May 27, 2009 10:23 AM
Have u tried praying for more powerful servers?
Posted by: Epinephrine | May 27, 2009 10:27 AM
Somewhat ironically, it wouldn't double post (told me I had submitted twice in too short a period), and I actually had to type it out twice. I just couldn't help myself ;)
Posted by: Fred the Hun
|
May 27, 2009 10:47 AM
Next parody apologies for parody double-posts?
We could play detect the very subtle differences from one post to the next...like change the double post font style to comic sans?!
Posted by: Standard Curve
|
May 27, 2009 10:51 AM
Maybe if you placed some crystals on the servers in a nice pattern, and said something really nice about Ohm's law...
Posted by: Larry Clapp | May 27, 2009 11:13 AM
Speaking of technical difficulties, there's a weird formatting error in some comments. See for example the comment starting "I do hope none of your family were with you when you typed that" on http://scienceblogs.com/pharyngula/2009/05/do_you_believe_in_angels.php , and several other comments on that page . The comment is out-dented for no apparent reason. (This is in Opera 9, Firefox/Iceweasel 3, and Konqueror 3.5.9, all on Linux.)
Apologies if this is a known bug. I'm new to your blog and ScienceBlogs.com.
Posted by: Richard Wolford
|
May 27, 2009 11:38 AM
Well, I am for hire if Scienceblogs needs yet another software developers :)
Posted by: tmaxPA | May 27, 2009 12:19 PM
It doesn't make any difference to some, but there are systems and browsers where it does.
My point is, you have to refresh the page, not just reload it. Reloading may make your browser go back to the last retrieved version, which doesn't have the comment. Refreshing forces the browser to re-retrieve the page from the server, which is what you want. There isn't a clear enough distinction made between these two in most cases, and so PZ's advice might actually be aggravating the situation.
Posted by: Pierce R. Butler | May 27, 2009 12:56 PM
Call in an exorcist who specializes in ping.chartbeats.net!
Posted by: Larry Clapp | May 27, 2009 12:59 PM
@ tmaxPA, #32:
If reloading the page doesn't hit the scienceblogs.com server, how could it aggravate the situation? Honest question; I may have misunderstood you.
Posted by: David Marjanović, OM | May 27, 2009 1:13 PM
Ever since the second-to-latest layout update, the first paragraph after a blockquote is edented unless you left a whole empty line below the blockquote (the way I do).
The latest layout update made the blockquotes themselves indented again (instead of slightly edented and thus hard to recognize).
By making people believe their comment didn't get through, so they post it again and again and again.
Posted by: Lurky | May 27, 2009 1:14 PM
#9 and #10
Hilarious :D
I can just about picture a theist going "NOW YOU SEE THE ERROR OF YOUR UNGODLY WAYS! GOD IS SMITING THEE!"
Somehow the best this most-powerful-being-in-the-universe-or-possibly-outside-of-it can manage against the evil atheists is a slightly slower performance and occasional double posts.
Posted by: llewelly | May 27, 2009 1:32 PM
I never bother to wait for the page to load after I click post. Even when scienceborgs is functioning 'smoothly' (well, as 'smoothly' as it ever functions), it's faster to use the back menu and reload.
Posted by: Ted Powell | May 27, 2009 2:30 PM
Anyone know what the underlying database software is? Could it be a product that requires the application to do explict locking/unlocking, and the recent code tweaks have subtly deranged this?
Posted by: 'Tis Himself, Quel Dommage
|
May 27, 2009 7:18 PM
I write a comment, post it, go to my desktop, play a hand of Microsoft Hearts, then go back to the thread and reload the current page. Every time my comment has been there.
Posted by: John Morales | May 27, 2009 8:22 PM
With some judicious tweaking of the hosts file, I've significantly sped up the loading of the page by getting rid of external crud.
Posted by: Nerd of Redhead, OM
|
May 27, 2009 8:34 PM
The posts seem to be submitted to SB fairly fast. A couple of times, I hit post, and after about a 5 count, I clicked back to the main page, and then after a couple of minutes, the post was visible in the recent posts column. The problem appears to be somewhere between the post submission and the page refresh. Something is getting hung up.
Posted by: Nerd of Redhead, OM
|
May 27, 2009 8:39 PM
Whoa, the above post went through like it used to... Trial 2
Posted by: Nerd of Redhead, OM
|
May 27, 2009 8:40 PM
Two in a row went through without the error message...
Posted by: africangenesis | May 27, 2009 8:45 PM
I post, then go do something else until the error page appears. Then I hit "back", then I select and copy my comment text, in case the post didn't take. Then I hit refresh, usually my post is there as it should be. Two or three times it hasn't been (out of many dozens), so I repost. one of those times resulted in a double post.
Posted by: raven | May 27, 2009 9:16 PM
That sounds good. Could you translate that into English?
I have third party cookies globally blocked. Tried to get rid of some "external crud" by blocking the websites i.e. sharethis.com. Nothing worked and some things seem to make it worse.
It does look like third party websites that add things to the page are slowing the loading down.
Posted by: George | May 27, 2009 10:07 PM
Your computer has been compromised. You are being moitored by government agents. I know because I also have experience weird computer operating characteristics even by MS standards. Ignor this at your paril! Watch your six and keep your powder dry.
Posted by: John Morales | May 27, 2009 10:07 PM
Raven, see here: hosts file.
Boils down to mapping hostnames to the local machine [127.0.0.1] (where they won't be found, and quickly too, when the page requests them).
There are premade host files redirecting known ad-servers etc on the Web you can download, or just look at the page source (or watch your browser when it says "waiting for [xxx]") to know what hosts it's getting data from.
Posted by: Alex Deam
|
May 28, 2009 12:53 AM
As long as they're not from Microsoft, you should be ok.
Unless that's been the problem all along? The Scienceblogs servers are run by Microsoft technicians?
Oh fuck, we're all screwed! Hell, Microsoft was Karl Marx's biggest influence. Das Kapital owes a lot to Bill Gates.
Posted by: raven | May 28, 2009 1:40 AM
What language is this now? LOL. I can barely understand what this means. I did however find page source and a long list which I gather is html web site design stuff. I blocked some of the obvious third party websites and things are loading quicker.
Thanks a lot. I've been using computers since the late 70's and never had a course in my life. I always know just enough to use them.
Posted by: Ichthyic | May 28, 2009 2:00 AM
or, you could try using adblock plus if you also use firefox:
https://addons.mozilla.org/en-US/firefox/addon/1865
if you're not using firefox, I'd recommend giving it a try.
Posted by: Britomart | May 28, 2009 1:04 PM
You know, as long as things are being fixed, can we have a better search function ? mebbe one that can sort ? or filter by date ?
And an RSS feed so its possible to get a notice when an old topic gets a new post. Or is that possible and I just don't know how to do it?
Thank you kindly
Posted by: Alex Deam
|
May 28, 2009 4:42 PM
Seconded.
Posted by: 'Tis Himself | May 28, 2009 6:39 PM
SB is getting wonkier by the minute. I did a reload to see if a comment posted and SB timed out.