If you’re seeing this post, you’re on the new server.
The old server isn’t around anymore. The CPU fan quit and the CPU fried itself. Luckily, we1 had moved everything over to the new server already and only had to do some final configuration.
I guess having the old server die completely is a good way to avoid the “blog comments on the old server while the DNS change propogates” problem.
While the web sites were down today, e-mail was down as well. If I haven’t responded to something you’ve sent, you might want to re-send it.
UPDATE: This affected FeedLounge too as the old server was serving DNS for feedlounge.com. The new DNS server settings should be propagating, but if you can’t access feedlounge.com try adding 65.90.218.228 to your list of DNS servers. Sorry about that folks.
:chuckle: I always thought that an enterprising plugin-writer would write a plugin that would do this for you: write the present comment_status and ping_status to a separate table, set comment_status and ping_status to closed, and use the copied fields to re-populate the table on the new server.
Of course, whenever I’ve made the move, I just did the move, disabled comments on the old server, and waited for DNS.
Why not just point the old website’s database settings to the new database?
Tienshiao – that would require the old server to have access to new servers’s database… which unless they happen to be in the same data centre (in which case why not just swap IP addresses) could be problematic.
And gah to feedlounge – going this long without access to my RSS feeds is criminal*. Plus what’s worse is this monthly payment went through today too. This is seriously making me consider changing back to a desktop feed reader.
* er, I don’t actually mean illegal… I just mean really really annoying.
Remember – FeedLounge is up (and has been this whole time). Just put the IP address above in as your DNS and you’re all set. We’re getting reports from users that the DNS change is reaching them and that they can access FL w/o using this DNS.
I think the problem is what you are asking a normal web user to do. Adding IPs to the list of DNS servers is NOT something 90% of users know how to do, let alone wknow what DNS is.
And not all members know to check your site if feedlounge is down.
I love feedlounge but you shouldn’t ask paying users to jump through hoops because you didn’t make it as easy as possible.
Dan, what would you recommend we do?
I’m not sure what you could have done differently, but I can say that I paid for feedlounge access and I didn’t have it yesterday. And I didn’t have access to any sort of status blog, or even this website. So that bugs me.
And it bugs me to see this comment
“Remember – FeedLounge is up (and has been this whole time).”
It’s at least academic to claim that the site was up when my.feedlounge.com did not give me my feeds. Especially when my money is already in your bank account. And don’t get me wrong, I like feedlounge (and tasks pro, for that matter), so I’m not just a grumpy user. But if I paid for access and I didn’t get it, please don’t remind that the site is *really* still up when I can’t access it. Please just apologize and promise to work harder to make sure it doesn’t go down again.
Maybe it’s time for someone to bring up the whole pricing issue again. 1 day of FeedLounge = $0.17. Are you more that $0.17 worth of “upset” that you couldn’t reach the service for a day? If so, then what a bargain you’re getting for the service!
I’m not Dan, but what I would/would have done would be to make sure you have two seperate machines running DNS, so that if one goes down, the other takes over. Two, and these days four, DNS hosts are pretty much the norm for any kind of 24/365 service. FL would have seen no downtime in the users’ eyes, and 17 cent comments could be avoided.
/vjl/
Yep, we had that. Unfortunately, the DNS set-up on the secondary server was changed a week prior to this and we didn’t realize it.
There is more on this on the FeedLounge blog.
More on the feedlounge blog many still can’t see. ;D
Vince: I’m still a noob sys. admin but that’s what I was thinking. Redundancy is key.
Alex: It urks me that this has happened because I have paid the premium for a premium product with premium service something you have givin me until this weekend. It’s just throughs me for a loop when I see something like this.
As for a solution: Other then what you’ve already told Vince, why do you not have a back running as we speak, a crappy 300mhz pc could do what you need to just do the redirection to the new server. Maybe I’m missing something but it seems like you just have the broken box sitting and just said “oh well”, “it’l fix itself”.
Another thing, as I said in the email. A lot of people do not know what the hell “The new DNS server settings should be propagating, but if you can’t access feedlounge.com try adding 65.90.218.228 to your list of DNS servers” means
Agreed with Bill with regards to “It’s at least academic to claim that the site was up”.
It was not available. Even if I had known it’s IP address, it wouldn’t have had the right cookies, thus it was unavailable. There is not getting around that.
And I’m sorry, but changing my DNS setup just for FeedLounge is unreasonable, no two ways around that either.
Let me tell you a story: I have two servers. On Saturday evening I rebooted one to get a new kernel, and it didn’t come back up. This is one of my DNS servers. While waiting for support to identify the problem*, I whipped on to the other one, updated DNS, and had a “Hardware failure, I’m working on it” page up for sites that were only on one server.
(the boot partition had issues. a new hard drive and the rest of the day copying configuration over).
Even more frustrating was that I tried to hit alexking.org, and of course that didn’t get me anywhere.
“Maybe it’s time for someone to bring up the whole pricing issue again. 1 day of FeedLounge = $0.17. Are you more that $0.17 worth of “upsetâ€? that you couldn’t reach the service for a day? If so, then what a bargain you’re getting for the service!”
This is the kind of cleverness that I don’t think many people appreciate in people they are paying. I’m interested in getting the best service I can, not getting the “best deal”.
Look, I run some websites, too. I understand that sometimes bad things happen and sometimes even good websites go down. But when I use a website, and when I pay for a website especially, it makes me feel better about problems when the people running the site at least apologize. Fundamentally, I’m not relying on the feedlounge code, or some set of machines, I’m relying on Alex and Scott (and whoever else might be involved these days) and they let me down.
Now, again, I’m not crying about this, or trying to be a bad user, or whatever. I like feedlounge and I like some other stuff Alex has done. My only point is that when I rely on a service and it lets me down, I feel better when that service indicates that it knows that it is at fault. Just a little “My bad, sorry, we’re gonna fix some stuff and it shouldn’t happen again” is enough. I don’t need an arm or a leg, or even a refund for 17 cents.
[…] From some of the comments on my previous post, I think I haven’t done a very good job explaining what happened and the nature of how this has affected FeedLounge. I’ll try to do this here, in the form of a O’Grady-style Q & A. […]
[…] The beta mailing lists are still down from the server meltdown, so please use the Beta Discussion forums to discuss these pre-releases with other customers. […]
[…] The beta mailing lists are still down from the server meltdown, so please use the Beta Discussion forums to discuss these pre-releases with other customers. […]