Site5 is an internet hosting provider that up until recently hosted several of the sites I manage. In 2013 my company had a project. This project required site hosting beyond what I was comfortable with providing in building as I was a new dad, hadn’t slept in months, and did not want a site to lose sleep over as well.
At the urging of a developer, we went with Site5 and had no issues for several years. There were one or two minor things that popped up, but standard and acceptable.
In 2017 they were sold, and everything went downhill from there.
In 2018 Pocketables made the move to Site5 as well. The unlimited storage, no bandwidth caps, etc seemed like the place to be. Site5 didn’t help with the move as it wasn’t coming from a Cpanel, but Vaultpress was helping do a restore over there and things didn’t seem particularly bumpy.
I think I had to open three tickets in a couple of days due to issues which were fixed within 48 hours… at that point just getting Pocketables moved seemed to be a gargantuan task and that it took a week didn’t seem particularly excessive. I had a lot of things on my mind with the site purchase and move from Crowdgather.
Later in 2018 we started trying to make Pocketables go faster. We were seeing 10 second load times per page, no CPU usage. I opened a ticket which ended up going through 14 reps if I remember correctly none of which read the ticket and just claimed that things were fine and they were not seeing the problem.
The problem was whenever our wordpress site looked up a site that had an IPV6 record the stupid resolution would time out at 5 seconds. When you’ve got plugins that work with other sites, this is a huge problem affecting thousands of Site5’s sites.
I very nicely gave them the fix, showed them how I’d determined the problem using the shell on Pocketables. I was informed that I didn’t have a shell, nobody did. Uh… I just cut and pasted results from your servers. I pulled a Karen and got to speak with a human finally.
Guy was in sales, but the only person I could actually speak with. I told him exactly what the issue was, gave him links to several sites that showed the issue and the fix, he agreed that looked like the problem but he was just in sales. Did know someone in the actual NOC but said it would require testing, could be months because they had thousands of sites and if it broke even half a percent of them they’d be overwhelmed.
He tried to upsell me on a managed VPS that would be about 12x the cost of the hosting we were currently getting and would have been significant out of pocket investment. I passed.
Two days later the fix was in place. The site was flying. I asked in the open ticket what had actually been done. I was told nothing had been done that it must have just been my site.
During this time the only way I could get a ticket responded to was to badger the Site5 twitter account and have them escalate it. Tickets would sit for days until I got a notification they were going to close unless I went and publicly asked them to work it, which then resulted in an escalation, then someone reading a tiny bit of the ticket and vomiting out an answer that usually was wrong.
This took months to get resolved.
On January 31st, for reasons unknown, VaultPress, the Jetpack service that backs up our site constantly, disconnected. Under normal circumstances this would trigger an email warning to me that there was a connectivity issue. It didn’t.
Should be noted, VaultPress got developers involved to figure out the issue and implement a fix which supposedly is now in place as of a couple of days ago. Simply bad timing.
On Feb 13th a little plugin I have called UpDraftPlus made a backup, shot that backup to an FTP server I’d forgotten about, and left a copy of the backup on Site5.
Feb 17th, sometime around 8am, an exploit for the Themegrill Demo Importer plugin wiped our site’s database along with 200,000+ other sites. I put in a ticket. Feb 18th, ticket’s not been touched, I asked them if I could pay someone to fix this.
I was told that the restore had completed. They hadn’t updated the ticket. Evidently they restored and it got instantly hacked. They updated the ticket and sent it back to administration.
They restored again. Said it was good. Tried to close the ticket. There was no content on the site.
Third time I started getting notifications that I was at the disk space limit and file number limit, which was odd because it was an unlimited account, and I had been only using a relatively small amount.
I was told the database would not mount due to there not being enough disk space. Uhh… ok… they said they had three backups for me, and needed to delete the local UpDraftPlus backup that was sitting there. The killed the upDraftPlus backup, restored three different versions of the site, and discovered that none of their backups worked.
Yeah, the three backups they had over the course of a month were all bad. Well, according to the tech of the moment. I believe I have 11 different people who worked this case. They’d deleted my backups.
They told me my database was wrong and I needed to fix it. They told me it was misconfigured and I needed to fix it. I was sent a couple of emails in which it appeared the tech was out of their league and was not bothering to look further.
I was asked if I had any backups. I said that I had the one they just killed, and was then given a link to the acceptable use policy that I was not allowed to have a backup stored on the server for any reason.
I was also informed that the backups they claimed to have were a courtesy to customers and not to be relied on. I was told this after they’d deleted the good UpDraftPlus backups.
That was that. I’d resigned myself to a 20 day old VaultPress version of the site until I discovered the UpDraftPlus backup had fired off to my home machine and was sitting there with no issues. We switched hosting and went live missing a couple of articles that night with the new hosts.
All sites moved off of Site5, I cancelled and was informed that we’re not getting our money back.
I’ll write a bit about the new hosts later. They got the site up and running in about 8 hours and did a bang up job.