After announcing the migration yesterday, the past 24 hours have been very stressful for me. About 11AM PDT yesterday, the migration began to move WUWT to a new much more secure and feature rich cloud server. It didn’t go well, and took far longer than I expected.
There were issues with the server SSL certificate, apparently, the SSL doesn’t become valid until DNS changes fully propagate. That was news to me, and the reason that some people got errors yesterday that the site “may be dangerous” because the SSL certificate wasn’t valid.
Propagation took an unusually long time, with the major issue being with Google’s public DNS servers, which didn’t get the memo until just before 10AM this morning. Now, once that log-jam broke, it appears we are essentially correctly referenced world-wide.
I have a long history with Google and repression. Me and our stable of guest authors write unpopular content, questioning the veracity of many outlandish claims about global warming. It’s inconvenient for some that we’ve been so successful.
Al Gore is on the BOD of Google was on the advisory board to Google, and in 2009 after I broke the story about ClimateGate he apparently lobbied Google to have me censured. Google abruptly and without explanation cancelled Google Adwords on my wordpress.com hosted site. Higher-ups at WordPress wanted to know why (they were losing revenue, after all), had a face-to-face meeting scheduled with Google the next week, and it was abruptly canceled the day before and they stopped returning calls.
My site’s treatment by Google is one of the strongest reasons the independent WordPress WordAds now exists, or so I’m told.
Through all this, WordPress and the WP staff (known as “happiness engineers” have conducted themselves with fairness, dignity, and without political prejudice. My hat is off to all of you, because you do in fact uphold the spirit of “net neutrality” while offering a superior platform. They don’t do, evil, and I sincerely thank them their help and professionalism through it all.
This is my first post on the new WP cloud server, where I’ve been since November 2007. God help me if I had chosen Blogspot (now owned by Google and called blogger.com) over WordPress.com back then.
Now, onward, and we’ll see if we can’t get a number of thorny little problems fixed while expanding features. Some comments made yesterday may have been lost, and remain on the old site. There is little I can do about that, my apologies.
Thanks sincerely to everyone for their patience.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Congratulations.
These things never go as smoothly as planned.
But a target achieved is a good thing.
First Murphy’s Law: Anything that can go wrong will go wrong.
First commentary to the first Murphy’s Law: Murphy was an optimist.
That’s why they call them “CONFUSERS”.
Bring back the slide rule and abacus !
Every task takes twice as long as you think it will take. If you double the time you think it will take, it will take four times as long. link
Mostly it’s that rascally factor of pi you forget, some times pi squared.
True but I would add three times the cost. A successful manager knows this but will not plan openly but keep it to themselves. On the money side, if they can, the budget stays closed to prying eyes. Saw this with a major ship acquisition and the time and budget were way over the contracted budget but came in under the managers budget. He made Admiral.
Just arrived here – was wondering why there hadn’t been a new post in the last day or two. Problem fixed now -thanks!
As for time estimates, my wife has a “rule” about my time estimates for getting things done around the house: “Double any number and then change to the next higher unit of time.”
This means that she allows 10 hours for a job that I say will take 5 minutes, and 4 days for something I say will be done in “a couple of hours.” This is based on years of observation.
In my defence, there is a difference in what we are talking about. My estimate is the actual working time, without allowance for getting the tools out, going to the hardware store to buy something I couldn’t find after looking for it for an hour, and then having a beer while confirming the final plan of attack in my mind.
Well, on balance, a huge move, and considering everything, not done badly at all. Some accidents are inevitable. You got away lightly. And well done for giving deserved credit to WordPress.
The like and vote buttons are gone?
Patience friend. Migration isn’t perfect.
Anthony,
If I had my druthers, ……. I would prefer that the new WUWT “cloud” server was transmitting ….. Times New Roman font and 11 point type ……. rather than the Helvetica font and 10 point type.
That 10pt type make for “slow” reading for an ole timer like myself whose eyesight is not that great as it use to be.
Sam C
That would be easier on the eye. At least the front is black; there’s a special place in hell for all those designers who use pastels for their fonts.
How do you get “pastels” in fonts?
Sam, if you use a Firefox browser – you do have a zoom feature that will automatically increase text size- that make it easier to read,
Yep… hold down the Ctrl key and hit you numerical keypad’s “+” sign. Works for most browsers.
Also, hold down Ctrl key and use the scroll wheel on your mouse.
Pretty sure it magnifies the page for all browsers/windows. I use it frequently.
I think you can also adjust (at least Windows systems) the computer to use your own specific font rather than what’s displayed in pages.
Have to support Mr Cogar on that one, Anthony.
There’s a very good reason why newspapers use a serif font — it’s called readability. And TNR is still one of the best. For anyone beyond middle age using an iPad this is just too small and too grey. Headaches will follow!
I’m over 70 and using an iPad. I just put two fingers lightly on my screen, and pulled them apart an inch. The text is now much bigger – magic, huh?
I’ve got mine CTRL+ up to 200% just to read comfortably. But the font looks like Times. It’s certainly a serif font.
I used to see this a lot with the old site, when the CSS didn’t always fully load – the font would remain a sans-serif and smaller. Often just a hard refresh would fix it.
Is there a way to adjust the font size as I and others have eye issues? AW you have ear problems so you should appreciate our difficulty.
Hold down the Control key, and scroll the mouse wheel
It’s not possible to adjust the font size as you’ve advised on a MacBook Pro. Could anyone help me?
Or, [Ctrl]++/- (Control key plus the ‘+’ or ‘-‘ key to increase or decrease font size).
“Hold down the Control key, and scroll the mouse wheel” doesn’t work on my MacBook Pro. Can anyone help me to enlarge and darken the font so I can read comfortably. It was not a problem for me before the change.
Command + to increse font size, command – to decrease.
I have found that increasing the page size by 10% will take care of most problems
My iPad does not have a mouse or control keys. I was content with the previous type and sized ee
Can’t you squeeze and expand using fingers? I thought iPads were famous for that feature.
None of these suggestions work on a MacBook Pro. I’m grateful that people are trying to help.
It looks like the font visibility issue has been resolved. Thank you. I’m sorry if this detail added to the stress of the transition.
@Sommer works on both of my macbook pros in all browsers to increase font size. I think is a windows-specific thing.
Murphey’s Law, baby. Murphey’s Law.
Love the new look and feel. Google is the anti Christ.
Compared to the UK’s TSB Bank migration, still not 100% sorted, it looks as if you’ve done very well. Great stuff.
Congratulations and well-done on the move Anthony. And I echo your thanks to Word Press for their adherence to professional and ethical integrity.
Perhaps the greatest surprise in the whole global warming/political thing is how readily, indeed how eagerly, so many people and so many institutions sacrificed integrity to politics. Living in freedom, they embrace tyranny.
Well, congratulations, Anthony. Actually you had an easy migration – I’ve been involved in much worse ones.
On Google and your DNS propagation, I’m not sure you can blame bias for that. I have noted that they are rather challenged outside of their core competencies – and running DNS is not one of those. Blogs run on the BlogSpot platform, whether left, right, or neutral are essentially cripples with the number of outside sites that have to respond before the simplest page is brought up.
Do you have data showing that Google is incompetent at managing DNS?
I had the problem yesterday, and am not going through the google public DNS servers. Also, I have some experience operating a web server with https, and DNS propagation should not produce what I saw. Failure to install the correct cert on the server would. DNS directing the browser to a server trying to do man-in-the middle could also produce what I saw, though that implies a server other than the old and new addresses for the migration.
I agree on perhaps not being able to blame google Anthony.
Every DNS registry entry has a reload time (TTL) associated with it. The default could be 1 day, or 1 hour, depending on who/when set it up… it’s best to decrease this to a smaller amount (say 30 minutes) a period greater than the previous value before you know you are actually going to change it… that way, when you do make the change it will propagate faster… then once it’s done you can again increase the TTL to it’s lazy old value… unless you can prove Google was taking significantly longer than the previous TTL (which you did not specify in the post), then claims of conspiracy are a little premature?
Prince Albert was a Google (now Alphabet) Advisory Board member in 2009, not on the BoD, as he was and is of Apple.
Not that the distinction makes any difference.
The site ‘may be dangerous’
Well Anthony a lot of people in positions of power have thought that for some time. Long may it continue to be so.
tonyb
So you use the Fallout test pattern?
With what somewhat looks like an Indian in “blackface”!?
Glad you are back.
TomB: The picture shown is the typical picture for technical difficulties.
Google’s search engine monkey business needs to be limited by legislation.
This is a science based site. Articles included data, refer to peer reviewed published papers, article assertions are supported by observations and logic.
No harsh name calling. The quality of the articles and the quality of the comments is one of the reasons why this is the best climate site in the world.
The problem is science does not support CAGW, hence the need for the cult of CAGW and their friends to try to shutdown climate sites that are independent and thoughtful.
P.S. Love the new edit feature. Best wishes all.
I think legislation to deal with Google cooking searches is a bad idea. It should be dealt with by means of anti-trust, as regulation would tend to make Google a protected incumbent, like Ma Bell was.
How do you “breakup” a “monopolistic” (*) search engine? What activities do you separate?
(*) monopoly, actor with too much market power, whatever
Making the company sell off Youtube as an example.
YouTube has too much market power now anyway. DailyMotion had its chance, they wasted it.
If Google wasn’t co-owned with YouTube, then what? That might level the playing field in a market where there is one big name…
Changing search engines is easy, once an offender is identified. Other online services (social networks) have network effect that make changing more difficult. But if certain pending EU legislation goes into effect (http://computingforever.com/2018/05/30/the-eu-is-about-to-destroy-the-internet-deleteart13/ ) it could do critical damage to any website that allows user comments and has legal exposure over there. I hope WUWT is a strictly US based operation.
“Google’s search engine monkey business needs to be limited by legislation.”
Limited HOW?
What exactly have they done wrong?
How is the Google search engine not free speech?
“Fallout” is new to me, we just called it the test pattern.
In its time the test pattern was highly useful. link In the early 1950s, test equipment was expensive and primitive. The test pattern was a godsend.
He’s referring to the Fallout game franchise, which frequently uses retro-1950s imagery.
As a youth I misspent many a late night with that test pattern on
after all three of the network stations in Cleveland went off the
air for the night…
Check out one of the final scenes in the movie Radioland
Murders where the patrol officers sit around mesmerized by
this test pattern.
It’s good to hear that Word Press is a company populated by people with principles. Duly noted.
Well done Anthony. A few hours delay is pretty good for these kind of complicated computery things. At least you didn’t have to try switching it off and then switching it on again.
You need to check the TTL on your DNS records prior to performing that kind of migration. Sounds like the TTL was set higher than you expected?
Yeah probably, but we changed nameservers entirely, and I thought that would override TTL setting. WordPress support was surprised too.
iirc any soap entry set to high ttl upon aws migration could cause this too. just been a bit so….I easily could be mike mann….I mean wrong.
however ….even WHEN changing nameservers set low ttl on both as (as you saw) it will cause some slowdowns.
man I have done way to many ns transfers LOL
Several levels of TTL here. The core nameservers — the ones that are authoritative for all subdomains of “.com” are one issue. Then the TTL on the records you publish through your nameservers are a second issue. All of these need to be considered when doing a migration. I get engaged frequently when “things aren’t working right” on migration night when if they’d called me the day before I could have prevented their problems. But once bad records are cached all over the internet there is nothing you can do.
So in your case, the core NS records are cycling between openhosting.com and wordpress.com. Strange but I get TTLs as high as a day on the core NS records. So if you change your nameservers, it could take up to a day before everyone is using the new ones. I’m not sure what you can do about the TTL and the core servers, if your changing them you need to be aware of that cache time and keep both old and new servers synced until the old core records time out.
The openhosting nameservers return a TTL of 1 hour on records in “wattsupwiththat.com”, which is reasonable. The wordpress servers also show a 1 hour TTL. What I don’t understand is how the core DNS servers seem to be out of sync on who is handling your domain.
In my experience, core DNS servers typically reflect updates at the registrar within minutes, so I should not be seeing what I am:
root@sun1:~# dig @a.gtld-servers.net. +norecurse wattsupwiththat.com ns
; <> DiG 9.9.5-9+deb8u15-Raspbian <> @a.gtld-servers.net. +norecurse wattsupwiththat.com ns
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 26761
;; flags: qr ra; QUERY: 1, ANSWER: 5, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;wattsupwiththat.com. IN NS
;; ANSWER SECTION:
wattsupwiththat.com. 62173 IN NS ns3.openhostingservice.com.
wattsupwiththat.com. 62173 IN NS ns4.openhostingservice.com.
wattsupwiththat.com. 62173 IN NS ns5.openhostingservice.com.
wattsupwiththat.com. 62173 IN NS ns1.openhostingservice.com.
wattsupwiththat.com. 62173 IN NS ns2.openhostingservice.com.
;; Query time: 19 msec
;; SERVER: 192.5.6.30#53(192.5.6.30)
;; WHEN: Wed May 30 23:22:08 EDT 2018
;; MSG SIZE rcvd: 157
root@sun1:~# dig @a.gtld-servers.net. +norecurse wattsupwiththat.com ns
; <> DiG 9.9.5-9+deb8u15-Raspbian <> @a.gtld-servers.net. +norecurse wattsupwiththat.com ns
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 53905
;; flags: qr ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;wattsupwiththat.com. IN NS
;; ANSWER SECTION:
wattsupwiththat.com. 13884 IN NS ns1.wordpress.com.
wattsupwiththat.com. 13884 IN NS ns2.wordpress.com.
wattsupwiththat.com. 13884 IN NS ns3.wordpress.com.
Querying the same authoritative core DNS server for ".com" shows two entirely different sets of nameservers which are supposed to be authoritative for your domain. With anycast DNS the "same" IP address could be many different physical servers but they are all supposed to be in sync.
Both sets of nameservers are providing the same address for "wattsupwiththat.com", so as far as I can tell this isn't causing a problem, but unless you're feeding both sets, eventually the zone will time out on the zombie servers (in about 14 days) and start returning NXDOMAIN (no such record). That will be disruptive.
I'd take this up with the registrar for your domain.
“I thought that would override TTL setting”
Which is my point, you have no idea how DNS works. Or any cache. The point of a cache is that records that are still valid won’t be checked by asking a server, so nobody will notice that the server is changed.
It’s fine not be an expert in everything. Many people here are not Internet experts, so they shouldn’t impugn the behavior of Google tools in that case.
I’ve done many migrations and if you can get it done in 48 hours, consider it a success.
As far as Google DNS. I doubt they purposely hung it up. It is generally a software managed process and it takes a while for a change to propagate out, especially through something as comprehensive as Google. If they have an issue with a DNS server, it may take them a while to diagnose and fix the issue.
But then again, your fears may be justified! I personally stay as far away from anything Google, from not using Android to using DuckDuckgo for searching and absolutely no gmail.
Thanks for everything, Anthony. People like you are rare.
Please describe the problems of “propagation” you have seen with Google Public DNS.
Like the sharp contrast of the website, very clean
Having spent 30 years before I retired implementing all types of IT system it sounds like it all went pretty smoothly, as it should in our modern day. Keep up the good work.
“Google abruptly and without explanation cancelled Google Adwords on my wordpress.com hosted site.”
It is almost like meeting Al Gore in person.
My usual thinking on any project is that it will cost twice as much and take twice as long as originally planned.
You clearly had the wrong Project Managers(;>)
David S,
That was my rule of thumb back in the days when I COULD work on cars. Now I let the dealer worry about it.
Good plan, that way it doesn’t take twice as long; it just costs twice as much.
Hmm… I’m seeing “Awards” twice along the top.
Anyone else?
Yeah, some sort of artifact, I’ll fix it. Thanks.
Slightly different look but still, much faster loading.
test comment
Yay! Comment editing is enabled! 5 minute timeout allows you to fix mistakes.
OMG!
Now I cn make a mistake and go back and fix it!
/grin
The only obvious problem is that yesterday’s post from Willis has disappeared.
No sooner said than fixed. Talk about fast!
It has been a long strange ride. Well worth the trip though. The new site seems very pleasant. Congratulations on a job well done.
The tips and notes page needs to be reversed in date order ie newest first and the Leave a Reply area at the top not the bottom of the page. Migration looks good