Problems, problems, problems...

[robb]

Originally Posted by eUKhost.com

You wont face any problems this time as we will move only your account and we will take care of verifying working of all your websites after transfer. You wont face any sort of downtime and you wont need to make immediate DNS changes as we will notify you once everything starts running smoothly from the new server.

I don't think it's quite as simple as that. Based on past experience, DNS propagation causes a couple of issues, once we make the DNS changes...

If a POP client uses the hostname then it may start collecting mail from the new server while some mail is still arriving at the old server - and such mail will effectively be lost.

On the other hand, if a POP client uses the IP address, this will need to be changed. If we wait until propagation has completed before changing the IP address, then no mail should be lost - but some mail may have been waiting at the new server for 72 hours.

To avoid losing or delaying any email, I typically access POP3 servers via IP address then create a new account for the new IP address when a site is due to move. That way I pull down mail that arrives on either server.

However, I had to to this for my clients when I moved to EUKHost. And I had to do it when my hosting was moved to EuroFighter. So I don't really want to have to do it to all my accounts and all my clients for the third time within a year...

Furthermore, sites with databases cannot necessarily run on both servers at once - as changes made to the database on the old server will be lost on the new server. (Of course, you could copy the database over, but then changes made to the database on the new server would be lost instead!) During propagation, site visitors may access either server and therefore either site and either database. To preserve data integrity, one of the databases has to be locked during propagation. This effectively causes downtime for those customers accessing that server. This is a particular issue for ecommerce sites as they lose around 50% of their sales during this time.

Is there any way to avoid these issues, or is some unavailability unavoidable for some site visitors?

[robb]

I know it has been Christmas and all that, but I feel like my problems have been forgotten about (e.g. no response to my last post).

In the meantime, it appears that Eurofighter seems to have been rebooted every day for the last few days, and only managed to stay up for a fortnight before that (http://uptime.netcraft.com/up/graph?...st.com&probe=1).

Furthermore, one of the problems that I regularly get with email has come back again (I can't email clients who host their web site with EUKHost, but not their domain name or email).

Not what I was expecting after EUKHost's promises...

[robb]

Originally Posted by robb

Furthermore, one of the problems that I regularly get with email has come back again (I can't email clients who host their web site with EUKHost, but not their domain name or email).

OK this has now been fixed. But I had hoped that things like this weren't going to materialise again...

And there's still the long term solution to be found.

[Rock]

Eurofighter has finally calmed down now. Few migrations were done off it.. Your account can be migrated too upon request... Open up a support ticket & it'd be done quickly..

[robb]

I don't want to move, I just want things to work!

I was given a number of promises at the end of last year, and lead to believe that things would be better in future. But instead, eurofighter has been even more unreliable than before and old problems have resurfaced.

I've also discovered that at least two important emails that clients sent me on 3 January completely failed to arrive! (The clients have different domains and are on different ISPs.)

From LiveChat at the end of last year:

NickG: I have already instructed our senior windows system administrators to watch your account.
NickG: We are confident you will find a marked improvement in the services & uptime
NickG: I am speaking to you in my capacity as the Sr.Manager - Quality Analysis.
...
NickG: I would request you to give us an opprtunity...

Well, I gave you the opportunity you asked for, but I haven't experienced "a marked improvement in the services & uptime". Quite the reverse...

Nick or "eukhost.com": any comment?

[eUKhost.com]

Hello Robb,

Please open a ticket for our windows team and ask them to move your account on some other server. Eurofighter had some problems till last week so that must have created problems for the emails sent by your clients. You can get your account moved on some other server if you are still experiencing any problems on Eurofighter.

You can specifically ask for Jack, Martin, Williams or Gary to look into your problem and they will get back to you with perfect solution. I personally have limited knowledge of Windows so we will need to wait till we get any reply from Jack or Martin in this thread.

[robb]

As I mentioned above, I can't currently see how moving to another server can be done without interruption to incoming email and sites with databases.

My clients went through such upheaval when they moved to EUKHost less than a year ago. They then went through it again when EUKHost decided to move all the sites I host with you from one server to another without any real notice - a move that was badly managed from both a customer relationship and a technical perspective.

So if I move again, it will be the third time within a year that clients have had to reconfigure all their email addresses and put up with at least partial downtime of their sites (i.e. by blocking access to site databases to prevent parallel updates).

While I'm sure I could explain that this is necessary to improve the service, and am also prepared to upgrade my hosting solution if that will give us increased reliability / manageability, my worry is that this will still not resolve our problems and I'll have to do it a fourth time - i.e. to move away from EUKHost. And that will be too many moves.

Part of the problem here is that no-one that I speak to at EUKHost will be fully aware of the conversations I've had with other people at EUKHost. Could one person at EUKHost take ownership of this scenario so we can look at pursuing the idea of setting up copies of the sites in parallel on another server (possibly under an upgraded hosting solution) and seeing what can be done to minimise the upheaval to clients (for example, whether we can move emails that continue to arrive at their old mailboxes to their new mailboxes)?

Based on my experience of support tickets (where different support technicians dip in and out, don't always read what has gone before, and have even done the exact opposite of what I have asked) I don't think that simply raising a support ticket will give us the holistic approach that we need...

[eUK-Nick]

Robb,

I will speak to our windows team specifically Martin, Gary & Williams. We will first set up a copy of your website on a different server, possibly one of our newest servers. Once that is done, I will get back to you. We can then take the process further from there. You may email me directly. I will have the communication done to the windows trio on your behalf.

Please allow me a few hours to have a parallel setup done. I will get back to you once this has been achieved.

We will do all that is required to have your accounts working correctly.

[robb]

Thanks Nick. I've sent you an email regarding that parallel set-up.

[eUK-Nick]

Robb,

I have now replied to your email

[robb]

Well, things have gone from really bad to even worse.

Five sites hosted with EUKHost have all failed for the same reason: the Application Pool crashing.

They are all low-traffic, the only other thing they have in common is that they are running on the same server, and one is just a single HTML-only web page. Yet, EUKHost claim that there's nothing wrong with the server! So how exactly can a single low-traffic HTML-only web page crash the Application Pool? If it can, then that means there's something wrong with the server anyway...

Three of these sites are still down as I write. One has been down for about three hours; the other two for about half that.

Separately, EUKHost have, in the last few days, compromised the security of all sites on this server to allow lazily written applications to run. I won't say exactly how in a public forum...

I have been hosting web sites for about 15 years, and I've dealt with some pretty poor hosting companies. But I have never seen anything nearly this bad before now. It has just been problem after problem after problem...

[eUK-Nick]

Originally Posted by robb

Well, things have gone from really bad to even worse.

Five sites hosted with EUKHost have all failed for the same reason: the Application Pool crashing.

They are all low-traffic, the only other thing they have in common is that they are running on the same server, and one is just a single HTML-only web page. Yet, EUKHost claim that there's nothing wrong with the server! So how exactly can a single low-traffic HTML-only web page crash the Application Pool? If it can, then that means there's something wrong with the server anyway...

Three of these sites are still down as I write. One has been down for about three hours; the other two for about half that.

Separately, EUKHost have, in the last few days, compromised the security of all sites on this server to allow lazily written applications to run. I won't say exactly how in a public forum...

I have been hosting web sites for about 15 years, and I've dealt with some pretty poor hosting companies. But I have never seen anything nearly this bad before now. It has just been problem after problem after problem...

Robb,

You have a support ticket raised on this issue to which Williams is replying. The issue of "security being compromised" is being replied to. Whatever apprehensions you have are being addressed.

[robb]

Originally Posted by eUK-Nick

Whatever apprehensions you have are being addressed.

They're really not.

Instead the ticket where I mentioned that three of my sites were down was closed before these sites were back up!

And I'm just being fobbed off with excuses that simply don't fit the facts, and contradict what I've been told previously...

[eUK-Nick]

Originally Posted by robb

They're really not.

Instead the ticket where I mentioned that three of my sites were down was closed before these sites were back up!

And I'm just being fobbed off with excuses that simply don't fit the facts, and contradict what I've been told previously...

Robb,

There is no reason for us to fobb you. The reasons provided are all valid & true. We do not believe in deceiving customers by giving false reasons. Your ticket has been responded to completely. All your 14 websites are functioning.

[robb]

Well, that's incorrect to start with. Two are STILL down.

[robb]

Secondly, you tell me that the application pool only crashes when the code uses a high amount of memory. However, there is NO POSSIBLE WAY that these sites have done so. As I have mentioned before, one of them is a (VERY) low-traffic, single-page, HTML-only site. How on earth can this have used so much memory that its application pool would crash? If you look at the logs you will see that this is clearly impossible.

You do not know why the Application Pool has failed for five different sites on the same server. Blaming the site code when that CANNOT be the problem is not the solution...

[eUK-Nick]

Originally Posted by robb

Well, that's incorrect to start with. Two are STILL down.

The domain names are working. I have attached the screen shots for that. The site preview is being fixed even as I write.

[robb]

Well, for one of them, the domain name doesn't (yet) resolve to your servers. So anything you see at that domain name is hosted elsewhere (and not the result of anything you're doing). That's why I gave you the preview address instead of the domain name. If the preview is down, the site is down, as there's no other way to access it...

[eUK-Nick]

Originally Posted by robb

Secondly, you tell me that the application pool only crashes when the code uses a high amount of memory. However, there is NO POSSIBLE WAY that these sites have done so. As I have mentioned before, one of them is a (VERY) low-traffic, single-page, HTML-only site. How on earth can this have used so much memory that its application pool would crash? If you look at the logs you will see that this is clearly impossible.

You do not know why the Application Pool has failed for five different sites on the same server. Blaming the site code when that CANNOT be the problem is not the solution...

Robb,

All your questions, doubts & issues are being diligently answered in the trouble ticket which you have raised. So there is no reason to keep on posting about eukhost not resolving it. Your posts seem to be so. Your tickets are being responded to very quickly.

The server would be slow during the early hours of Sunday owing to the backups being taken. There is no run-around for that.

[eUK-Nick]

Originally Posted by robb

Well, for one of them, the domain name doesn't (yet) resolve to your servers. So anything you see at that domain name is hosted elsewhere (and not the result of anything you're doing). That's why I gave you the preview address instead of the domain name. If the preview is down, the site is down, as there's no other way to access it...

This has also been addressed in the ticket.