Leaving OpenVZ.ca/Media-Hosts.com 2019-07-09

Back in the late summer of 2015 I found DNSCrypt, which is basically a protocol for wrapping DNS queries in encryption. I started playing with it and in the early fall I made a post about it. Back then I started with okturtles and a few European servers but I had occasional name resolution troubles. I fiddled a bit but eventually just went back to my ISP's resolvers since they just seemed to always work. Some time passed and I found myself wanting to go back to DNSCrypt, and since there was no specifically Canadian DNSCrypt presence I decided I would try running a pair of my own resolvers for people to use. I tried out a few VPS providers (CloudAtCost was a total mess, and IntegralHost had crappy performance) and finally settled on OpenVZ.ca, which is just a sub-part of Media-Hosts.com.

It was actually late September 2017 when I added the resolvers to the public list, and the amount of traffic it used was significantly more than I had expected. I opened a ticket on 2017-10-01 with the media-hosts.com support asking for bandwidth forgiveness. 17.5 hours later I got a reply from Joseph Conti saying:

Your bandwidth usage isn't at a level where we would be concerned. There is no additional charges at this point. Unless you're going over by hundreds of gigabytes we usually don't bill overages.

Clearly I was happy about this as I watched my bandwidth climb in to the hundreds of gigs and the folks at media-hosts.com never complained or said that they would have to start charging me for overage. This went on for 20 months. The bandwidth fluctuated some over that time, but was always hundreds of gigs, month after month. The service was also pretty stable. Occasionally a process would just silently die though. So I wrote a script that would check the dnscrypt-wrapper processes every minute and would restart them if they suddenly disappeared. Life was good. Then on 2019-06-13 I received an email telling me that there would be a maintenance window on 2019-06-22 from 10AM to 8PM. I was disappointed that such a long window would essentially mean the resolvers would/could be down for most of the day but I was optimistic that the change would help to improve their infrastructure. Then at 4PM, in the middle of the maintenance window I received an email saying the window would last an extra two hours until 10PM. As it turns out, one of my servers came back up at 12:30AM the next day (2.5 hours AFTER the expected end of the maintenance window) and I fell asleep waiting for the second one which came back up at around 4AM (six hours AFTER the expected end of the maintenance window). That sucked, and I was pretty disappointed. No email, no explanation, no nothing. Just downtime.

I stole this pic from DraftKings.

It went downhill from there too, on 2019-06-30 I received an email telling me that I owed just over $86 for my bandwidth overage. No warning, no email about a policy change, just a bill. A bill that seems pretty out of touch with the $15 per YEAR I am paying for the servers. Assuming that the overage would be the same each month that is basically telling me that my annual cost for the two servers is going to go from $30 to $1064.16. There is of course the option to upgrade to a larger package which would drop the cost enormously, however... it is still clearly a dick move to just hammer me with a giant bill with zero warning. I immediately opened a ticket with them asking if I could return to the previous state of bandwidth forgiveness. I was hoping that the change was the result of some billing system work they had done weeks earlier and that the agreement we had in place for 20 months could continue. 11 minutes later I received a reply saying:

Thanks for reaching out,

I have our Billing Administration team reviewing this issues and as soon as they are able to provide a follow up with you regarding this.

They will be sure to keep you posted.

Reach out anytime should you have any further issues or questions.

... but I got no answer, so a day later I replied to the ticket with:

Bump.

... and got no reply. Another day passed and I replied with:

Could I please get a reply on this? I am already past my limit for July.

At this point it was clear nobody cared to answer me, so I went looking for another VPS provider. In the meantime the servers have been running but I had to remove them from the public resolver list since they are just using up bandwidth without me having any idea what is going on with Media-Hosts. So in short, the servers are still there but my services are down because I can't get an answer from Media-Hosts.com. Finally this morning I sent an update to the ticket saying:

What the actual fuck?

Two minutes later I recived this reply:

I'm not sure what this is in regards to since I am seeing no forwarded message.

Was there something we can help with?

Which is odd, because the entire replyless ticket was there to see. This was followed by another reply eleven minutes after that saying:

I have cancelled the overage invoice as a good gesture. This is a one-time offer since it is your first overage invoice

It is not clear to me what has happened to OpenVZ.ca/Media-Hosts.com and I don't know why it takes more than a week to get a response to a support ticket or why I need to write a "WTF" message to get a reply and I don't know the motivation behind them deciding to hammer me with billing with zero warning. Maybe it is related to the recent announcement that cPanel is pissing off hosting providers with a huge price increase. Maybe someone died and the person responsible for their policy management is a douche. Maybe I am a dick for thinking that it would be nice of them to warn me of a massive price increase/policy change before hammering me with a bigass bill. What is clear though is that OpenVZ.ca/Media-Hosts.com is no longer a viable hosting provider for dnscrypt.ca or the two other VPS's I rent from them. Frustrating and diappointing to say the least.