This is becoming a joke

[DaveTheShark]

My business hosted site has been down yet again for the last 10 minutes.

This is really becoming a joke.

[Tony]

Did you open a ticket? What was the response? What site is it anyway - I'm wondering if I can get to it from here.

[DaveTheShark]

Yes, also live support. First response is.
"your server was experiencing a temporary problems with MySQL. Our technicians has successfully resolved the case."

Server uis now back up.

This is the second time I believe mysql has crashed the server in a month.
Suspect someone isn't using indexes on their database.

The server has been running for 10 minutes and look at this data

400.6 k	The number of requests to read a row based on a fixed position. This is high if you are doing a lot of queries that require sorting of the result. You probably have a lot of queries that require MySQL to scan whole tables or you have joins that don't use keys properly.
44.2 M	The number of requests to read the next row in the data file. This is high if you are doing a lot of table scans. Generally this suggests that your tables are not properly indexed or that your queries are not written to take advantage of the indexes you have.

[DaveTheShark]

I emailed this question to support. No reply.

[Stas G.]

Could you please specify what e-mail address did you use to contact our technical support? If we receive an e-mail to some of our departments you always get an automatic reply with ticket ID. If you received one please provide us with your ticket ID and we will check it.

[DaveTheShark]

I just replied to the email I received from support.

[Stas G.]

There is a ticket ID in the subject of every e-mail you receive from us. It contains 3 letters and 6 digits divided with the dash. Could you please provide us with it, so that we could check it?

[DaveTheShark]

Sure, PNV-728419

I included an html page I'd copied from mysql stats showing all the stats.

[Stas G.]

Dave,

First of all, we did respond to your e-mail. According to the ticket you’ve mentioned, the latest reply was given by our representative. In case you didn’t receive it we recommend you to login to our helpdesk website at https://www.whbsupport.com/index.php with your e-mail address and check the ticket in question. So no lack of customer support was experienced. For the future please keep in mind the initial maximum response time on all our tickets is 60 minutes, and there is no need to post on the forum that you didn’t receive a reply from us at once after the ticket is submitted. It took 10 and 9 minutes correspondingly for our technicians to reply to your posts in this ticket.

As for the server you are hosted on. You have a business account with us, and there is only one difference between shared and business servers in resource means: there are fewer accounts on business servers. For example, there are 59 accounts on your server, and that’s quite a good value (taking into account we advertise up to 100 accounts at our website). The server with your account has 8 CPUs Intel Xeon E31230 3.20GHz and 16 Gb of RAM, and all these resources are shared between all accounts on the server. There can’t be an isolated environment on shared/reseller/business server due to their structure and the needs they are intended for. So you should always keep in mind there are other people using the same shared resources.

As for the specific issue you have faced with. There was one account on your server that started consuming lots of server resources, therefore the server overload occurred. Unfortunately, such problems can’t be prevented on servers shared by their nature between many customers. Your issue also has no relation to possible overload of the server by new accounts that we may setup there all the time, according to the numbers and specifications stated above. We have taken all appropriate actions, suspended the abusive account and the server got back to its normal performance. As our technicians confirmed, no server issues with MySQL service have been detected, there was just an overload related to the intensive use of server resources by one account.

As for the uptime of your server. The uptime is considered as the period of time when the server is operational. On the other hand, the downtime is the period of time when it’s not operational at all. Temporary issues with availability of separate services (mail, web, FTP etc.) are not considered as downtime, because other services are running properly at that moment. So your most recent issue is not counted as downtime. However, according to our monitoring system the current uptime of the server your account is hosted on is 37d 12h 39m 7s. Indeed, there was an issue with your server in October specified in the forum branch dedicated to your server. Somehow this incident was not counted as downtime, though it should have been, so it’s our fault and we sincerely apologize for this misunderstanding.

Summarizing all mentioned above, we recommend you to upgrade to VPS if temporary unavailability of web services affects your business greatly. The main advantage of Virtual Private server is the isolated environment, where you will have your own amount of system resources assigned to your server only and you will be able to distribute them as you wish. I would like to emphasize that we recommend you the upgrade not because your account is incompatible with business server, but because the resources on such servers are shared between all users and load issues may arise from time to time due to irresponsibility of some customers on the server. But I can assure you when such issues occur we always face them with all our experience, knowledge and skills and try to solve them as soon as possible.

[DaveTheShark]

Dave,

First of all, we did respond to your e-mail.

OK, yes you did but I sent 2 emails. First asking if you'd resolved what the problem was or just restarted the service. The second I'd sent you a copy of the stats from MySQL which looks like it is being very heavily worked. There are a lot of lines in red that highlight areas that may kill performance. The email I received from you said "The account, which was causing this issue was suspended and MySQL service was restarted."
I assumed this was in my response to the first email I'd sent as the 2nd one crossed in the path. If you check the log you will see that they were both sent at 19 Nov 2011 04:46 PM. This email cannot have been read prior to your response. I therefore maintain that my 2nd email remains unanswered.

there is no need to post on the forum that you didn’t receive a reply from us at once after the ticket is submitted. It took 10 and 9 minutes correspondingly for our technicians to reply to your posts in this ticket.

I posted the comment 16 hours after sending the email.

I understand what you say about moving to VPS but I do not want to do so at the moment. The server seems to run fine, perhaps only using less then 10% of its cpu most of the time. The odd amount of down time is acceptable but I don't understand why a business service is causing me more issues then my previous £15 a year provider. Don't forget the real headache for me over the last month has been the network problems. VPS would not have cured this.

Finally, this is the 2nd time that MySQL has died, was it the same problem? same Account?

[Matt R.]

Hello Dave,

As I've posted before, I take the level of service we provide very seriously and noted you've posted a few times regarding this.

Firstly, let me clarify that our business hosting is also a shared environment. Other users can, unfortunately, impact on the performance of the entire server. That's just the nature of shared hosting that you will find with all shared hosting companies. A virtualized or dedicated environment removes the ability of any one user being able to impact the server for everyone else.

Secondly, on your ticket submission:

Ticket submitted: Posted on: 19 Nov 2011 04:21 PM
Staff reply: Posted on: 19 Nov 2011 04:31 PM
Your reply: Posted on: 19 Nov 2011 04:37 PM
Your second reply: Posted on: 19 Nov 2011 04:46 PM
Staff reply: Posted on: 19 Nov 2011 04:46 PM

You had to wait a maximum of 10 minutes for a staff reply. Where does 16 hours come in? As my team said, you have to include the ticket ID in the subject for it to get parsed by the ticket system.

Third, onto the server itself.

business3 is a SandyBridge Xeon, the latest type of Xeon from Intel. There are currently only 59 accounts on the entire server. Business servers generally get 100 accounts, regular shared servers generally get about 4-500 accounts. Our competitors regularly load 1000+ accounts on the same spec servers we use for 50% that number. The load is low, performance looks great right now and I even pulled a bandwidth graph of November (attached). As you can see, bandwidth has been going in/out at all times. A complete outage would show up as a complete blip in the graph.
graph.png

At this point, I'm not sure what else we can do. If you wish to close your account and move elsewhere, we will of course accommodate. If you wish to move away to a non-shared environment, we currently have a £1 first month offer on our Silver VPS on WebHostingBuzz.co.uk (I see you are in the UK). The coupon code to active this is vps30daytrial. We can, of course, assist with the transfer of your accounts should you wish to do this.

Matt

[DaveTheShark]

I'll answer your points in a bit Matt.

Meanwhile the server is yet again unavailable.

[Matt R.]

Dave, no it is not. I've been personally logged in since I wrote that post.

[Matt R.]

System information
Server load 0.24 (8 CPUs)
Memory Used 15.63% (2,572,748 of 16,458,380)
Swap Used 0.01% (992 of 8,388,600)

[DaveTheShark]

The server is now available.

Suspect the on-going network issue.

I have sent Matt and suport a copy of a report from HostTracker [Website Uptime Monitoring Service]

This shows 50 odd networks tests from around the globe, Most of them show a very long page load of about 20 seconds.

If I'd run it a bit earlier I'm positive it would have been even worse.

Matt, I was sat on my pc in the UK and was unable to connect to the server for 2 minutes.

[Matt R.]

Dave,

I'm sat at my workstation also in the UK and have been logged into the server continuously. I'm watching the bandwidth (network) usage along with server resource usage. Bandwidth usage is normal, resource usage is low. Any network issues would have resulted in a closed SSH session and a decrease in bandwidth usage.

[DaveTheShark]

Which server Matt? My server and whb server go different routes. I've noticed this before, my server will be unavailable but I can still reach whb.

[Matt R.]

Dave,

Business3. Both are in the same network and location as the main sites by the way.

Matt

[DaveTheShark]

Matt, Did you check out the link I sent you via PM?

Notice how the average response was 18.7 seconds. It's now 0.5.

[Matt R.]

I just saw it now. Is there any reason you are using third party / your old hosting company's DNS? It could be them timing out and not us. The HostTracker information doesn't make sense with such a wide range in results from hugely diverse locations.