Log in

No account? Create an account
Website outages - MoonScape [entries|archive|friends|userinfo]

[ website | My Website ]
[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

Website outages [Apr. 16th, 2015|05:20 pm]
[Current Mood |annoyed]

For those of you who visit any of my websites or read the site-embedded blogs...yes, they're all down.   It's Verizon's fault.  They had some kind of problem night before last that destabilized my hosting service for over six hours...and by the time Verizon claimed they were up and stable again, they had done "something" than knocked the hosting service and all its sites down.  First they swore they'd fix it yesterday.  No.   Then they announced that someone would be "in" at 8 am today and that was the right team to fix it.  No.  Then in the afternoon they sent an emal to the hosting site to say the right people would be in late tonight.   I am not holding my breath.

As my hosting service says, "You could not make this up."    This is not the first time Verizon has knocked over the china shelves, so to speak, at my hosting service.

But if you have noticed the sites are down--yes, I know (have known since before the "fix" when the connection was off-and-on) and yes, I know it's not the hosting service, but Verizon.   It will get fixed eventually, although the desire to pour a bucket of wet fish down the back of Verizon's CEO's jacket is becoming stronger.  (The ability to do this is of course limited by common sense and better things to do.  Like use the time I can't be blogging on the websites to, um, work on the current chapter.

From: geekmerc
2015-04-16 10:52 pm (UTC)
Web host's fault. They should have had a more redundant setup to avoid this type of problem. Even BGP flap dampening penalties can be avoided through monitoring and action. :)

I mention this, as it is something that should be considered in the future if you value better availability of your websites.
(Reply) (Thread)
From: geekmerc
2015-04-17 03:38 am (UTC)
Yeah. Very bad design. Your web server and both DNS servers are in the same IP Address subnet. In fact, it's in a /25 assigned to GREYWARE AUTOMATION. This is bad, as it is too small to advertise via BGP, which means it is locked into verizon only; I'm guessing it's a Fiber To The Premise connection to your web host.

Now, this is not necessarily a bad issue, except that apparently they aren't running a mirrored setup to a different location and IP Address. In addition, sff/paksworld has a TTL (time to live) of 72 hours. This means that even if they did try and change the IP address to a backup data center IP address, it would take up to 72 hours for all caching DNS servers to update.

I'd have thought that SFF was on a higher end product. :(

I mean, a low TTL could have meant spinning up a virtual at digital ocean and putting a capture all webpage that said, "We're sorry, but we are temporarily down." and redirecting all the domains to it. It's only like $10/mo for something simple like that and it's actually billed in sub-hour increments. It also would have let people know what was going on instead of them getting browser connection errors.

It's like a high speed insertion. Without problems, it isn't a big deal. If problems arise, you have no maneuverability to escape disaster.

Edited at 2015-04-17 04:52 am (UTC)
(Reply) (Parent) (Thread)
[User Picture]From: ckd
2015-04-17 05:50 am (UTC)
Yeah. There's a reason that Google, LinkedIn, Facebook and other similar companies have a role called "Site Reliability Engineer"; there's an explicit focus on "what does it take to keep this service up and running in a reliable fashion?".

Sounds like sff.net could use some of that in their network design, alas.
(Reply) (Parent) (Thread)
[User Picture]From: e_moon60
2015-04-17 12:48 pm (UTC)
And there's a reason this thread wasn't titled "Why I hate my hosting service."

Enough already.
(Reply) (Parent) (Thread)
From: geekmerc
2015-04-17 02:02 pm (UTC)
I'm sorry. I suspected that you liked the web host. Looks like a small local Texas company. I was just trying to provide you additional information. My second reply actually has useful information that your web host can use to adapt for the future.

I apologize that I was poor in my wording. Rereading it, I was a little harsh. Even though I registered the people aspect, I ignored it in my technical analysis. A poor technical design isn't necessarily reflective of the person managing it. Costs, size and experience play a role. From what I've seen in the past, the actual site management looked great, but I'm ignorant when it comes to website design and configuration of the common elements that compose the back end of the website. My expertise lies in networking, availability and the base servers themselves.

I think I've mentioned before, my social skills aren't the best. They are at their worst when I do technical analysis. I focus too much on the technical and forget the business and social aspects. Anyways, sorry.

Edited at 2015-04-17 02:18 pm (UTC)
(Reply) (Parent) (Thread)
[User Picture]From: e_moon60
2015-04-17 02:35 pm (UTC)
Understood. Greyware gives excellent service in many ways--communicates with customers, has extra services included which I appreciate. And they were incredibly patient and helpful with people new to the internet at all. My sites are down very rarely. Having had sites split up among different hosting services before, I had a chance to compare and did not like the others. They had nothing to do with my website design (that's from Willowbrook, a small design service in which I used to be active for graphics until I got too busy.)

There are multiple considerations in choosing whom to go with on a project; having had "irreconcilable differences" with others, I am indeed happy with Greyware/SFF.net.
(Reply) (Parent) (Thread)
[User Picture]From: redvixen
2015-04-17 04:33 pm (UTC)
Hmmm, yes, that would be insulting to the fish.

Hope the issue is cleared up soon and in the meantime you find something more enjoyable to do, like maybe more knitting or going for a walk. :)
(Reply) (Thread)
[User Picture]From: e_moon60
2015-04-17 08:07 pm (UTC)
Or...mess about on Twitter, here at LiveJournal, or the many, many other sites I have bookmarked... Or I could write on the book. I have knitted today because it's gray and gloomy and big storms are moving this way at a variable speed that makes going out to the far corners of the place iffy. The clouds here are heavy enough that I can't tell by looking where the "severe storms" are w/o checking radar online.

And it's not like I have no indoor work to do...
(Reply) (Parent) (Thread)
[User Picture]From: redvixen
2015-04-17 11:09 pm (UTC)
Lol, there's always indoor work to do. It keeps coming back. :P

Hope you did enjoy your day, whatever you ended up doing.
(Reply) (Parent) (Thread)
[User Picture]From: e_moon60
2015-04-17 11:20 pm (UTC)
Sock #2 of the Summer Beach shorty pair continues to give me trouble. I'm not happy with the toe; I didn't start the decreases in the right place, apparently. And I've had foot cramps today, which means curling my foot around in a convenient way to try on the sock and see if it's ready for another decrease has led to some more foot cramps and less accurate measurement, as my toes and foot "curl" when cramping.
(Reply) (Parent) (Thread)
[User Picture]From: e_moon60
2015-04-17 11:53 pm (UTC)
I'm afraid I need to rip it back five or six rows, but not until tomorrow when my feet aren't so sensitive.
(Reply) (Parent) (Thread)
[User Picture]From: redvixen
2015-04-18 01:59 pm (UTC)
Ugh, foot cramps are painful. I get them when my magnesium is low so I try to avoid that state. Especially since extremely low magnesium can be fatal. I've been too close to that state and never want to go there again.

Well, hopefully you can fix the mistake with minimal ripping out. At least you don't have to start over with it.
(Reply) (Parent) (Thread)
From: (Anonymous)
2015-04-20 02:34 pm (UTC)

ripping out

Rip out the socks, not the feet.

I am glad to know you are relatively all right - I started getting frightened when your website went down.

(Reply) (Parent) (Thread)
[User Picture]From: e_moon60
2015-04-20 03:03 pm (UTC)

Re: ripping out

The most likely cause of a website going down has nothing to do with the website owner but with the connectivity, whether it be a guy with a backhoe on a new road project severing one (usually) or two (uncommon but it happens) big fat fiberoptic lines or a Bigname Company deciding that customer complaints from Smallfry aren't worth the bother. (SFFnet is far from the only customer to be affected by this outage. As of today, the complaint finally reached the Director level.)
(Reply) (Parent) (Thread)
From: ozdragonlady
2015-04-20 07:51 am (UTC)
mmm .. . still out today? Im not getting to either blogs or your own page.
Checking the traffic report I see that one of the Texas gateways isnt functioning ....

For what its worth, to a degree I agree with the advice offered. Sometimes loyalty has to be reconsidered in the face of economic necessity. You have a very large following and are known world wide. Access to your pages is part of the marketing process.

Looking forward to the new book :)

Edited at 2015-04-20 07:56 am (UTC)
(Reply) (Thread)
[User Picture]From: e_moon60
2015-04-20 03:05 pm (UTC)
I'm well aware of the marketing importance. This is not just about personal loyalty but about service over time, with services that I value highly and other web hosts do not provide reliably.

And anyway--I've made it clear that it's my decision, and it's made. If I change it in future it will be on my own terms, for my own reasons.
(Reply) (Parent) (Thread)