Website down? "Apache" error? Need to order! :D

curiousswede · Post by **curiousswede** » Wed Jul 18, 2018 7:24 pm

Hi! I was trying to go to your eu.dollstudio.org to order a M16 magnetic connector but I was met with "Apache" error as if server was installed with new php code or something? (see image below)

Anyhow, I hope I can order your M16 magnetic connector (32mm version) ASAP! Feel free to contact me. Do you send it from Germany or China? It would be awesome if you have it in stock in Germany for faster delivery, but I completely understand if not.

Forb · Post by **Forb** » Wed Jul 18, 2018 7:33 pm

I'm not an IT guy but that looks like the web server has just finished updating or they've reinstalled the OS. That might take awhile if it's not a mistake.
I would try emailing them, or PMing them here on TDF.

Post by **Dollstudio** » Wed Jul 18, 2018 7:52 pm

Hi,

yeah, since everybody seems to be on vacation lately I have started to set up a fresh server. Moving the data will take a couple of days, about 120 Gig/day.

DNS updates might be propagated faster than the data, from a few hours to 48 hrs. Depending on which resolver you query you will be either pointed to the old or the new server.

Sandro

Post by **Dollstudio** » Sun Jul 22, 2018 7:28 pm

Hi,

I ran into some (more or less) unexpected issues with the migration.

There is no estimate possible when the sites will be back up again.

For your 'entertainment': This is what happens if OVH is responsible for you DEDICATED ROOT SERVER:

HARD Reboot
Date 2018-07-23 00:25:26 CEST (UTC +02:00), maxence C made HARD Reboot:

Actions:
Restart the server hardware.

No notice in advance.

These people are just braindead… HARD REBOOTING a dedicated root server while I'm logged in and working on it…

Sandro

Post by **Tsall** » Mon Jul 23, 2018 2:38 am

Maybee you should migrate to another hoster? I never had anything like this with my dedicated root server hosted by Hetzner. They always inform in advance.

Post by **Dollstudio** » Mon Jul 23, 2018 3:12 pm

Hi,

in a nutshell, this batshit crazy technician keep hard resetting the server all night long, at least four times. Then he started operation "Electric connector", meaning he replaced the PSU. Everything without advanced notice. Today I got a new server, so I am now juggling data between four root servers and basically start from the beginning (as of July 18th).

Regarding alternative hosters, there are not so many good options if you have bare breasts on your site. Check the T&C for restrictions regarding inappropriate content. Diapers yes, sex dolls not so much.

Sandro

nukeno · Post by **nukeno** » Tue Jul 24, 2018 9:47 am

I remember that you had similar problems last year or so and I was all worried and started a thread like this one...

If this is becoming some sort of annual tradition, you could run at least a backup server at home or at your office.
Maybe not as fast as a "professional" solution, but cheaper and it keeps you running at least.

TDF members know and trust you, but anyone from outside this community, who visits your page thinking about buying a doll
one day, but can't reach your site the next day, doesn't think twice and buys somewhere else...

Post by **Dollstudio** » Tue Jul 24, 2018 1:12 pm

Hi,

Code: Select all

sent 243,883,302,226 bytes  received 77,483,713 bytes  15,829,788.53 bytes/sec
total size is 257,627,180,055  speedup is 1.06

Nice

So at least the data has now moved, again. Now back to where I was on 18th…

@nukeno: Running an active backup would require failover servers and database replication. That's what one would do with a seriously flourishing business. The little 'side effect' of such a topology is that you can not just add 'one' backup server, you need to run a small server farm including stuff like a load balancer and spares for everything. So the next step would be (at least) 6 instead of 2 servers which is not too realistic.

Sandro

nukeno · Post by **nukeno** » Tue Jul 24, 2018 1:59 pm

Dollstudio wrote:@nukeno: Running an active backup would require failover servers and database replication. That's what one would do with a seriously flourishing business. The little 'side effect' of such a topology is that you can not just add 'one' backup server, you need to run a small server farm including stuff like a load balancer and spares for everything. So the next step would be (at least) 6 instead of 2 servers which is not too realistic.

No, you are right, that would be overkill.
I don't think that mirroring the main server is necessary in your case. Your business is not that critical.
However, just a simple html server, nothing fancy, no database etc. one you could easily run at home
or on your business laptop, so you can keep a store front with your logo and contact information up
as long as the main server is down.

Post by **Dollstudio** » Wed Jul 25, 2018 12:14 am

@nukeno: You mean something like a site maintenance page… that needs to be where the DNS records point to (= the server).

I think my reply above wasn't properly clear. The downtime is not caused by a technical failure, that was only the 'topping', generously donated by OVH last night. The root cause for the migration is simply an EOL (end of life) issue of several components. EOL software is no longer maintained and not supported with security updates, so it's just too risky to keep running an outdated system. Updating was unavoidable, and a time without orders is a good occasion to deal with this troublesome stuff.

What is biting me now in the ass are other components which are not available in the new environment (insignificant stuff like displaying description texts on product pages), and other things are 'too new' and not yet supported (other insignificant things like Varnish caching). And this will go on in the next months; Debian has already dropped support for MySQL, Drupal (8) will drop support for PHP 5.x branch next year. On the other hand, D8 considers migrations from older versions still 'experimental', and the shop system is still an 'Alpha' release since 2016. Somewhere in-between I'm right now.

At the moment I am not entirely sure if I will put the sites back online. The way they render at the moment can be admired at us.dollstudio.org (no text descriptions display, no caching). In the next migration step I will loose the theming (how the site looks) and a lot of other things dear to me, too…

Sandro

nukeno · Post by **nukeno** » Wed Jul 25, 2018 7:31 am

Dollstudio wrote:In the next migration step I will loose the theming (how the site looks) and a lot of other things dear to me, too…

Yes, a simple maintenance page. That's much better now, then getting 404s.
I assumed the server was dead or the service unreliable because you'd talked about hard resets earlier.

You seem to be into this technical stuff, luckily I quit trying before I got sucked into it to deep.
Not because it is too complicated to do yourself, but it's so nerve and time consuming...

Best wishes!

harddk · Post by **harddk** » Fri Jul 27, 2018 4:38 pm

Since i currently cannot use your homepage to keep track of my order, I'll ask here.
Are you making use of DHL Express and with the company name 'BCD International Co.' when mailing dolls?

Reason being, I've just recieved notification from them about an expected delievery, but not sure if it's the doll or some accessories I've ordered. The message and tracking no. doesn't mention weight or size.

Post by **Dollstudio** » Sat Jul 28, 2018 2:34 am

@harddk: Your doll will travel with UPS. I sent you an email with further information.

Please note that the tracking code will become active after she passed customs. Unless the backlag has cleared until then, you might be affected by this:

https://www.dollforum.com/forum/viewtop ... 7&t=102815

Sandro

Post by **Dollstudio** » Sat Jul 28, 2018 4:36 pm

Hi,

the websites are fully up again. The issue with text descriptions not being displayed has been resolved; it was caused by a different PCRE handling of the newer PHP version.

Varnish is not yet enabled, so the sites will probably respond slower.

So far I'm not aware of other regressions. If you come across an issue, plese send me an email at info@dollstudio.org.

Thanks,

Sandro