Terrible support and response time.

[Xinil]

At 7:30am PST today, my site MySQL database was removed from hostgators servers. No e-mail, no phone call, nothing to let me know of this. The only way I knew was by going to my site, and having a nice big "cannot find mydatabase.frm". Wonderful.

At 7:37am PST, I submitted a ticket requesting to have this fixed, and to explain why my database was removed. Two minutes later (wow), I received a response telling me my database was suspended

Dear Garrett,

It looks as if your database was suspended, I'll forward this to the abuse queue so they can find out why and let you know.
Daniel O.
HostGator
System Administrator

Three hours later, 10:16am (my site is still offline), I get an e-mail from support with:

Hello,
We have been forced to disable the database xinil_maldb on the gator229 server. The queries running on
this db were causing high server loads. Please review our logs below and let us know what can
be done to prevent further issues.

There's also a whole bunch of info below that which details what queries were being run at the time of the shut down.

Considering the fact that I've been running those same queries for the past 2 months with no problems, I didn't understand how I could be using 70% CPU resources. I replied to their ticket (at 1:10pm) and asked what could be wrong with my queries and I also mentioned that they were doing fine for the past two months.

No response.

2:03pm I send another e-mail requesting for help in figuring out exactly what I should be doing to fix this. I can't even access my database to perform MySQL explain statements.

No response.

3:00pm I log on to Live Chat and ask for help to have an admin check my account. I'm told an admin will check my Abuse Ticket soon.

Insert many live chats, phone calls and other junk here. Each rep told me my ticket would be looked at soon, and I should be patient.

8:25pm PST, no one has replied to my ticket yet. I go back to Live Chat and ask what the heck is up. The rep tells me that I must EMAIL abuse@hostgator.com to get my database re-enabled. Wow? No where did anyone tell me that I had to e-mail support. This rep also kindly told me that I needed to specifically say what I would do to fix the problem. I understand completely that I was supposed to send an e-mail explaining how I would rectify my query, but I wanted more information about the problem before I changed anything. No replies came back from those.

8:30pm, I send an e-mail to abuse@hostgator.com explaining that I"ve rewritten my SQL query and it should be fixed. Funny, e-mailing abuse@hostgator.com just adds the e-mail to the end of your support ticket. Doesn't do much in terms of escalating the problem.

It's now 11:30pm, my site has been down since 7:30am. I have not received one response about re-enabling my SQL database. Absolutely ridiculous support. I've spoken to HostGator by phone and by live chat multiple other times today too, and each time they told me they could do nothing and I would have to wait for someone in abuse to check out my problem.

Very displeased.

[Xinil]

My MySQL was re-enabled this morning and my databse had the tables reinserted again at 7:30am. 24 hours down time total. I don't blame it all on hostgator, as it's obviously my fault it originally went down, but I'd say at least 18 of those hours are waiting for Hostgator to respond.

At 12:25pm PST my database was shut down again. I find this incredibly odd as I was monitoring the CPU usage all morning. I changed my queries and I changed around sites and added explicit "mysql_close()" commands on all my pages. I am nearly positive that my site did not exceed 25% CPU usage over 90 seconds (as it says in the ToS) so I'm very confused as to how my site got disabled again.

While monitoring the server CPU Usage earlier this morning (when my site was online) I noticed that CPU usage would randomly spike to 30% (even when I disabled my site myself). So, obviously there are other users on the server who are taking a load off the computer.

I'm pretty sure I narrowed down the problem with my changes/additions, but I'm positive I could fix it if HostGator would allow members to change their mysql settings. Current MySQL wait_timeout is set at 3600 seconds. That's quite high. 1 hour just to timeout a sleeping connection? If you set that down to 60 seconds it'd make a HUGE different in the number of processes running.

It's now 1:50pm, so my site has almost been down for 2 hours. I have yet to receive a response from Hostgator about my second downtime too. We're at 26 hours downtime in about 32 hours total. My site grows in size and "boom" throttled...great.

Please read the rules of the customer review forum before posting. This has been moved.

[Xinil]

Apparently I exceeded the CPU load limit at 11.4% CPU usage. Wow, that's impressive.

Here's the message I got the first time my account was suspended:

op - 09:26:59 up 81 days, 8:30, 1 user, load average: 11.00, 12.07, 11.62
Tasks: 273 total, 3 running, 269 sleeping, 0 stopped, 1 zombie
Cpu(s): 70.5% us, 7.4% sy, 14.7% ni, 6.3% id, 0.0% wa, 0.0% hi, 1.1% si
Mem: 4143580k total, 3962264k used, 181316k free, 167920k buffers
Swap: 6289408k total, 4320k used, 6285088k free, 2739352k cached

70% CPU usage, that makes sense that it was suspended.

And now, here's the second message I just got when it was suspended:

top - 14:12:39 up 82 days, 13:16, 1 user, load average: 11.08, 13.12, 13.05
Tasks: 267 total, 6 running, 260 sleeping, 0 stopped, 1 zombie
Cpu(s): 11.4% us, 10.3% sy, 26.6% ni, 49.2% id, 2.1% wa, 0.0% hi, 0.4% si
Mem: 4143580k total, 3887224k used, 256356k free, 136752k buffers
Swap: 6289408k total, 112k used, 6289296k free, 2813224k cached

Suspended at 11.4%. How is that warranted?

[twisty]

Are you using persistent connections for mysql? That will get you in trouble every time.

[Serra]

Originally Posted by Xinil

Suspended at 11.4%. How is that warranted?

You are looking at the overall CPU load, you really need to look at the load of your process. Do you have the rest of the email?

Also, databases get larger. The same queries that were fine 2 months ago may take longer and use more resources when the db has increased in size.