Why I resist offering my usual levels of support for issues related to domain purchases and management.


I make my profits on web hosting. The domain names are essentially a loss leader for me; it costs me about as much as you pay when you buy a domain from me, when you take into account credit card processing fees and the like. I don't mind when customers buy domains; they often turn into web hosting customers. It's a fully automated process, and rarely requires any kind of work for me aside from refilling my account from time to time, or upgrading the software.

However, sometimes customers decide that they want me to do normal domain management tasks for them. This in spite of the fact that the domain management tools I provide are easily accessible and not hard to use. They reason that they don't know the system and want to get it right, or that it'll only take me a few minutes, or that I owe them something for being a customer, or that somehow my personal, manual intervention in something they should be able to do themselves is part of the service.

In fact, I often end up doing that kind of routine work for customers. However, over the years I've noticed that since my domain prices are so low, many customers buy domains without buying a web site, so I have literally thousands of domains in my system for which I do not earn a profit. I found that more and more often I was spending time working with these domains, which is literally a money losing proposition for me, because my time is my main commodity. I am therefore quite resistant to offering much in the way of support for domains, because the customer is supposed to be able to handle most tasks themselves, and a few minutes multiplied by thousands of domains means I lose my shirt.

I have no problem whatsoever in helping customers with non-routine tasks related to domains. With web hosting, I even provide help with routine tasks, of course. But I am simply unable, for the reasons mentioned above, to give routine help with domains. I would go broke. And this is why I resist and encourage customers to use the system I have provided for them. Because my only other options are to either raise my prices and offer higher service levels or to go out of business.

I ran this past a friend who is in the business, and he gently suggested that I "ease up on the snark", because it's my choice to charge the prices I charge. He's absolutely right. I could charge more, have fewer customers, and make fewer conversions from domain sales to web hosting sales, have less business, and be less stressed over this issue. It's just that when I've spent a lot of time on this stuff with people who aren't hosting sites with me, it's hard to make the connection; and I don't want to raise prices on that basis because that essentially punishes customers who I'd like to keep. Which takes me back to the original point; the reason it can be so cheap is because I should have to do very very little to help with these domains. I am not complaining about customers who genuinely can't cope with managing their domains; I am complaining about customer who should be able to do so but would rather I do it.

Long time no post.

qm2 is safe for now, I have bigger fish to fry.

Been working a lot with external cheap VPSes, results are mixed but not awful.

Also running a trial of Directadmin on what is currently called web4 but should probably get a dwarf name because I don't anticipate making it a H-Sphere box. Reason: I'm getting sick of Ensim on grumpy, the last upgrade to PHP using ART repository was really a disaster. So I was thinking (with Vam's help) of doing a two step migration from grumpy to a cpanel box, then to directadmin.

Directadmin's pros include

• Vam likes it (which is big, because he's being very helpful in sysadmin support),
• it has a lifetime license (plusses and minuses there)
• is cheaper than cPanel monthly/yearly even w/o lifetime license

Of course, I already have the Ensim license, but the installation is so hosed on grumpy at this point that I don't think I can keep using it much longer, and unless they suddenly start supporting it, keeping sites there will eventually cause a big fail, which no one wants.

I have replaced the 256 GB drives on the "web4" box with two 1 TB drives, and even though I cannot add to the RAM short of replacing the MB, it should be okay. I had Vam install DA there. We did a couple of test migrations, and they worked, but there are some missing features, especially the inability to add jailed FTP subaccounts automatically via the control panel.

Also, reading the DA forums leads me to believe the DA team has less depth than the cPanel team. While I like to hack around as much as the next guy, I just don't have the time, and I (perhaps naively) believe I'll have less to do if I choose cPanel. My current thinking is to install Xen on the box, and then make one honking big VPS inside running cPanel. This is because Xen itself doesn't have a huge footprint, and cPanel for VPS has a lower footprint and license fee, so the tradeoff probably works. Big question is, can the box support all the sites on grumpy? Rambling: it's probably about time I learned cPanel, no?

whew

The "bashful" server isn't on next month's invoice! YAY! Still have qm2/3, working on a plan for that. So far have moved DNS off. Or did I say that already?

and more stuff

• Some PCI issues
• win2008 on aboud
• upgrade h-sphere (again, yay, they're maintaining it)
• free up hardware "just in case" (yeah, right)

bashful bites the dust this month

No, really.

It was up until yesterday ns.alsonetworks.net and ns3.alsonetworks.net and it had some other random info on it. I ended up getting three geographically disperse cheap-o VPSes, total cost around $15 and used webmin to clone the DNS stuff on bashful to the new "ns" VPS, and slaved the new "ns2" and "ns3" VPSes. Then I changed DNS A records and re-registered the name servers. After a 24 hour wait, I took down named on bashful, and checked. All is well. Then I converted "ns" to master, and violin's your uncle.

So I sent to softlayer and canceled bashful. It goes down on the 18th.

Next project: drop qm2 / qm3 on softlayer. If the VPSes on burst work out, there's no real reason to keep even a cheap dedicated on softlayer. The qm2 / qm3 server functions as DNS, nagios and pbcs backups. As mentioned above, DNS is now off that box (I did it at the same time as I did bashful). Creating a new Nagios installation will be a bit tedious but will give me an excuse to investigate cool front ends (opsview?) for the tool. Once that's done I'll probably have to upgrade the disk on one of the other VPSes (no reason to keep Nagios and backups on the same box) to accommodate the pbcs backup stuff.

The VPSes: two in USA (Scranton, LA) and one in EU (Manchester). Nicely disperse geographically for DNS. 512MB, 20 GB HDD. Took a bit of tweaking, especially Manchester, but they seem okay.

Can easily put Nagios on one. EU or USA? Don't know, hafta think.

Backups won't fit in 20 GB, choose either to expand or just get a different VPS. Also hafta think.

Anyway next project is Nagios on somewhere outside of qm2 / qm3.

sheesh

Boy, have I let this drop.

At least I now have a "direction" for moving the DNS. The question is, where? I also have to archive a few things, and grab Squirrelmail address book and aliases. That's about it.

no more sites on bashful

...but DNS is proving to be a pain in the neck. Over 100 domains to adjust. I've set up the cluster to receive the DNS but I seriously doubt (with the way things are going) that I'll get it cleared before the billing deadline.