“I have been a customer with LeafyHost since June of 2004 (back when they were Aetheri) and at this point I’m very fed up with them. I had great patience with them when they had their hard drive failure last year because they had been a very good host for my site. I use my domain largely for e-mail and hosting images, so downtime wasn’t really an issue for me. Also, Christopher and I had been posting on the same forum (not Ars) for something like six years and he hooked me up with a free hosting package after my prior host ( vectorstar.net) had massive downtime issues because they refused to replace a failing network card (absurd, I know). They had been prompt in handling what few support issues I brought to their attention and all was well. Things have changed for the worse since the drive failure. At this point LeafyHost is on yet another server because “our current colocation provider has been having problems with their upstream bandwidth providers, and combined with the issues with securing more power in the past that they’ve had, they are moving out of the data center where we are currently located.” This was handled with few hiccups with the exception of my site being a 404 even after all the dust had settled, but as far as I can tell that wasn’t their fault. The issue had resolved itself by the time I fired off an e-mail to them.

The problem I have with them at this point is that I have two support tickets that have been open for several months. No updated information for the tickets has been supplied in about as long as the tickets have been open. The first ticket is a real minor issue that I just brought to their attention since it was affecting all users (DSPAM died). I was told on 2/22/07 that “It looks like we have a fix, but we need to test it out through our dev server first. Once that looks to be stable, I’ll move it over to Uffish and update this ticket.” Nothing since. It’s a minor issue since Thunderbird has done a superb job of filtering out most of the junk mail I receive, but I am frustrated with the lack of *any* information since February.

The second ticket is a much more severe issue for me: I can receive mail, but am unable to send mail. The last piece of information I received on this issue was when Sam told me he was passing the issue over to Chris and that he would look into it. I haven’t heard a thing from either Chris or Sam on the issue since that day. I have sent several e-mails to them over the last two months asking if any progress has been made on the issue and have not had a single reply to any of them. No “Hey, we’re still looking into it,” or a “We’re working on fixing it now,” or even a “Fuck off.” Nothing. Complete and total silence. After the migration to their third server the whole ticket system is down. You go to the page and get a page saying that they need to install the Zend optimizer on the server before the page can be displayed.

This is complete and utter bullshit. A web hosting company should respond to a user’s requests for information pertaining to the quality of services being provided. They should respond in a prompt manner with as much information as is available at the time. Before the site went down with the migration a few other users had been complaining in the LeafyHost forum about the lack of communication from Chris and Sam, or if they got a reply it was slow in coming. At this point I have to wonder if Chris and Sam even care about their hosting service at all. Have they decided that since things appear to be spiraling down the toilet they might as well piss off their customers? Have they decided that no news is good news? Are they playing NetHack all day?

These guys need to start taking their business much more seriously if they hope to make it through 2007.”

I am not instantly discrediting them. This post was made after months of false promises. If they would’ve admitted that they messed up and then fixed the issues, things would have turned out differently.

Unfortunately they chose to handle the situation by pretending to fix the problems, then running away for a few days. Some time later they would come back and the whole process would start over again.

(Note: Sorry for the delay in getting your comment up. I was away for a couple of days and didn’t get a chance to approve it.)

On another note, what WOULD be a suitable backup? I don’t know anything about hosting, so I’m just curious.

Bzzt. Wrong. Please try again.

Drive one. The one that went bad.

Drive two - “Since we weren’t syncing deltas and instead were just copying the data from one drive to the others, once the IO errors started the other drive copied over its corrupted data.” Therefore this wasn’t a valid backup.

Drive three - “The initial power outages knocked that out.” Therefore it too was not a valid backup.

The truth is, 90% of our users had their data recovered immediately and suffered several days of downtime, which never, ever, should have happened.

The other 10%? That’s the 10% we really screwed up with, and our initial hope was to make things as secure and stable as possible for that 90% so that the majority wouldn’t suffer because of an issue with the minority. That’s the largest lesson to learn, and now that we have, we’re doing our best to make happy the 10% as well, and to recover as much data as possible.

“Although they advertised multiple backups it become very obvious that they were a false promise. There were no backups.”

That’s absolutely not true and it’s unfortunate you’re passing it off as a truth. We had three hard drives, one of them live, one of them mounted in the server, mirroring the first, and one of them remotely available. Since we weren’t syncing deltas and instead were just copying the data from one drive to the others, once the IO errors started the other drive copied over its corrupted data.

As for the second drive in the server? The initial power outages knocked that out, believe us, we tried to recover from backups, it just wasn’t an option, and not for lack of hard drive redundancies. Lesson learned? Keep a drive or tape in a safe.

We had a trial license and were considering Helpspot when the server had it’s explosion, so we stuck with it. A few weeks after The Event, the trial license expired. We asked for an extension, and were granted one. Recently, that license ran out, and now we are considering our options. The guy behind Helpspot has been incredibly gracious with us, we are just looking in to all possibilities right now. Helpspot is definitely still one of them.