[QuadAC]

Update:

The server is now up and running.

After talking to UK support on Monday, the server was scheduled to have its disks replaced, without much evidence that they were actually at fault. I was informed that I should perform a hardware diagnostic, which found out that the memory was actually at fault. Once this was replaced, and one more re-install performed, the server is up and stable.

Some observations:

Why wasn't I informed that I could do a hardware diagnostic last week, when talking to support?

Why are support unable to instigate this themselves? I had reported that the server was dead, multiple times, yet the only response was to re-install it - are the technicians in France incapable of initiating the hardware test themselves? I understand this is an unmanaged service, but they work 24/7 and kicking off a non-destructive hardware diagnostic on a dead server shouldn't be beyond their ken.

Why was there a mass of miscommunications between the UK and France, with me on the outside making what I thought were perfectly reasonable requests? The attitude of 'it lights up, so it's up to you to fix' on an allegedly brand-new server doesn't seem to be reasonable to me.

The server was ordered in English, was re-installed (multiple times) in English, and yet was installed in French with UI set to English, requiring the language to be changed post-install. Surely an English image should be installed in English from the get-go? A minor point but again points to some sloppiness in implementation.

Why wasn't the faulty memory detected when the server was commissioned? It took over a day for the server to be made available after purchase, why wasn't the hardware checked during this process? It only takes 20 minutes or so and would have saved everyone a lot of grief.

All of this has made my boss understandably wary of trusting OVH with anything critical, and also feels that it took the threat of a credit card chargeback before we received a decent level of support. However, I personally cannot fault the support we received this week; however the ball was dropped completely last week. We shall see how everything performs from now on, and we will be seeing what they come up with as compensation for this mess as per their SLA before committing to keep the server.

[Winit]

Performing a charge back isn't a good idea. OVH may cancel all servers if they detect a charge back. The same process is often used by fraudsters.

[Andy]

Update:

The server is again failing to boot, a technician has again looked at it and again it is reporting that the file \WINDOWS\SYSTEM32\CONFIG\SYSTEM is missing or corrupt.

And again they put it into KVM mode in spite of my mulitple requests to replace the entire server.

And so it goes on...

If they won't replace the server, purchase a new one and do a charge back on your card for the current one. Thats probably the quickest and simplest solution. Either that or wait until Monday for support to return and phone them about the problem. OVH's weak point is e-mail support and weekday 9-6 support, never rely on it.

[QuadAC]

Update:

The server is again failing to boot, a technician has again looked at it and again it is reporting that the file \WINDOWS\SYSTEM32\CONFIG\SYSTEM is missing or corrupt.

And again they put it into KVM mode in spite of my mulitple requests to replace the entire server.

And so it goes on...

[QuadAC]

Here is a tale of woe.

We ordered a server on the 3rd of December, one of the 64 bit Windows dedicated server, with a little indicator saying there were four available within an hour. Payment was accepted and we received the server the following day (slightly longer than an hour, but no great hardship).

As soon as I remote desktopped into the server, it reported that it had recovered from a serious error - in other words, it had blue screened. Not an ideal start but I soldiered on.

After turing on IIS (it is, after all, going to be a web server) the desktop became corrupted, and then locked up. I initiated a remote reboot but it failed to come back. Cue the first call to technical support.

The chap at the end of the phone was curtious enough, however he simply said that I should raise a ticket, which I did, and a technician would look at it.

After a couple of hours, I recieved a reply indicating that the server had come back in KVM mode, and how to connect to it. This I did only to see that Windows was unable to load the system registry hive - basically Windows was unable to load at all. I called support, who again said I should raise a ticket with the support peoplem, however by this time it was obvious to me that the server itself was at fault and suffering a serious problem. No matter how many times I asked for the server to be replaced I was told that there was nothing to be done and that I should wait for the technician to look it over. Needless to say I was not impressed by this response, so I asked them to raise the ticket and make sure to say that I thought that the server was suffering from a hardware failure.

A reponse was received from support person Angie, who, in a very poorly worded message said 'please reinstall the server'. I had no idea that *I* had to do this, I thought this was sent to another person in the data centre and I was being cc'd into the email for information. So the server went un-reinstalled.

The following day, with the server still dead, I sent another email to find out what was happening. I was then told that I had to start the reinstall - the server was brand new and dead, surely it should have occured to someone that it needed reinstalled at once?. So I started the reinstall and waited.

And waited.

And waited.

It took over 4 hours to install, and then failed during the verification stage, and promptly started installing itself again without my requesting it.

So, I was back on the phone to support, to a reasonable man who coped admirably with my ranting about how this was completely unaccepable. He informed me that I had to wait for the reinstall to finish before I could get it looked at again by a technician, the fact that it took over 4 hours last time meant it would be evening before it would finish. On a Friday, when I was no longer at work. The support person said that it wasn't unheard of for a server to fail installation once, only to success the second time. To say I find this incredulous is putting it mildly.

In the end it took only an hour to reinstall. When it promptly failed again with the same error. And started installing itself, again.

Third time was the charm, it passed verification and I logged into the re-installed server, just before I had to leave work, full of hope.

The screen corrupted, complained of running out of memory (on an 8Gb 64 bit Windows install), then locked up.

Precisely the same problem I saw in the beginning of this sorry tale.

Right now I'm looking at a support ticket indicating the server is fine, it is pingable and the services are running. However I can't connect to it. I have no confidence in it. Simply put, it is broken.

So now it is Saturday, three days since ordering the server. It has never worked for more than five minutes, it is constantly crashing, and I've recieved terrible support from the get-go.

If this isn't fixed before Monday, and I don't mean reinstalling this completely broken machine, I mean supplying a new server, and making sure it is fully functinoal, we will be informing our credit card company that this service wasn't acceptable and performing a charge back.

In my (many) years in dealing with dedicated server suppliers on both sides of the Atlantic, this is, by far, the worst handled support problem I have seen. By far.

If you have made it to the end of this lengthy post I thank you for your time, and I would tell you think carefully of putting any critical work soley in the hands of OVH - I don't think over three days downtime due to obviously faulty hardware means they live up to their SLA, do you?