2020-08-05: Caveat emptor: OVH hosting                       rak

After three years hosting my primary VPS at OVH, I migrated my
VPS to DigitalOcean this morning. Why? Because of 42 hours of
continuous and unexplained downtime, significant data loss, and
poor customer service.

My VPS became unresponsive around 18:10 on August 3rd: it did
not respond to pings, I could not access its console through
OVH's web interface, I could not reboot it, etc. I filed a
ticket, which they escalated at 23:59. I (politely) followed-up
in writing at 9:21 and 14:54 the next day asking if they had any
updates and never heard back.

My ticket was getting me nowhere, so I tried their phone support
around 16:10. After sitting on hold for 25 minutes, I was told
by their customer service representative that because I only had
"standard" support, that there was nothing they could do until a
"technician" looked at my issue, and that I should sit tight
meanwhile. I was also told that they couldn't process any refund
requests until the issue was resolved.

Shortly after noon today, about 42 hours later, my VPS came back
online. The catch? It appears to have been restored from a
snapshot made around 5:50AM on July 14, 2020.

Over the past decade, I've hosted VPSs with several providers,
often for 4-5 years at a time. I've encountered unexpected
downtime, but in each case, the provider clearly communicated
the issues with me and the downtime lasted at most a few hours.

Fortunately, I run several VPSs at different providers for
redundancy, so I didn't lose any mail or important data. My
setup is also well-documented and my deployment is mostly
scripted, so all I lost was an hour or two this morning spinning
up a new VPS at DigitalOcean (and much of that was spent
restoring backups and waiting for DNS to propagate so that I
could get a new SSL certificates from LetsEncrypt).

I should have listened to the many reports I read three years
ago that OVH's customer support was horrible. Lesson learned.
Caveat emptor.