This post is a follow-up to this earlier one and is second in the series. You really do need to have read the first to understand what’s going on in this oneI’m afraid.

As the big day for the Lotus Domino Migration approaches a few changes to the current infrastructure have to be made, and obviously a few checks have to happen.

Now I’m not going to get into the grubby details of change control management, it’s a horrible topic full of mindless boredom, so instead I’ll stick with the interesting (for me anyway) stuff. For the rest of this article, and all others related, please imagine that both Change Management and Due Diligence has in fact taken place.

To test the feasibility of the plan outlined in the last post three things would have to be true. First is that the IP changes we’re planning on would have to work. Secondly, the version of Domino on our eventual combined server is up to date and stable. Thirdly, and possibly most importantly, will migrating actually work – and if not what is the backup plan?

Luckily, we use VMWare ESX with Virtual Center in my company which makes testing immensely easy, which means we actually managed to successfully test two of the three issues highlighted above, implemented one in production and came up with a solution to the third.

Firstly we had to looked at the upgrade from Domino 6.5 to 8.5. Specifically whether it could be done in place, whether the updated databases were still useable and to ensure that resource usage wasn’t increased by too much. This was probably the simplest part of my day. First, I simply took a clone of my current Domino server through VMWare, booted it and downloaded Domino 8.5. I started the process at 8am, and by 9am I’d run the setup program and installed everything over the old installation. It was then simply a case of letting Domino boot, which took about 5 minutes, and then make all the database and mail file changes it needed to so that we were bought up to date. These database and mail file changes took another 30 minutes. But by 10am we were routing test mails through the dev server and everything was working quite nicely. It even turns out that, even though it loads up as four times as many processes, overall Domino 8.5 has about a 25% reduced memory footprint overall from 6.5 which was a nice bonus.

Due to it still only being 10am, and the test deployment going so well, we decided to make a snapshot and do the update live. By temporarily giving the host more Memory and an extra 2 CPU’s, we managed to have the whole process complete with only an hour’s overall downtime, and at the client this was even less because they could connect before their database had been updated (little did they know they couldn’t use it though).

Next on my test list was actually migrating a user. Now that we have a nice new version of domino running, we thought we’d try the documented way and find the user in the address book and “move user to a different server”. After 3 hours of clicking, scratching and reading logs I gave this up as a job gone wrong so decided to look for a different way of doing this, and then it came to me. Instead of migrating people with a single button, I decided to try Replicating the mailbox from one server to the other, and then simply updating the address book to show that the server hosting the mail has changed. Replication took some time, because of the distance between sites and the slow WAN, but after it had finished all I had to do was change one address book entry and replicate that throughout Europe (forcefully) and suddenly the migrated user is working like a charm, with all changes being replicated still to his old server!

The next thing to test was whether the IP differences between the old server and the new when migrated to my site would cause issues. It was decided that instead of running a mixed network we’d just give the old mail server a new IP address and replicate that throughout the company. We’d immediately ask the offsite mail users to use my mail server instead of theirs, and slowly migrate routing responsibilities there too. This would mean that for a time they’d be using a secondary replica of their mail, but as we could replicate every 5 minutes this isn’t a major issue as far as I can see.

So, where does this leave us. With little over a week left until we actually make these changes only one additional thing has to happen over the above. We have to give the hosts mail server a bigger HDD to cope with all the extra data. Again because we’re using VMWare, the provision will be simple, we’ll literally just assign more space to the database drive. When combined with 2008, which can allocate additional diskspace on the fly, I should be able to assign another couple of hundred GB with no downtime on the server at all.

All we have to do now is prepare the offsite users by telling them to setup new Mail connections, and start the migration. We can move all the smaller databases live over the WAN in the week to come, and then worry about the big databases come the Weekend of the move.

There will then probably only be one more blog post on this move, and that will be the one detailing the actual move.

Since becoming the de-facto URL shortner for Twitter bit.ly is taking its position seriously. The service has teamed with StopBadware.org and AntiPhishing.org to help prevent users from accidently clicking on a link which contains possibly harmful material for your PC.

Although this isn’t a new idea, safe.mn have been doing this from launch, with the audience of bit.ly and twitter combined, this can only be a positive step in protecting us from otherwise unknown dangers within the links we see on our feeds.

Google has launched a tool to migrate Lotus Notes users from Domino into Google Apps. The Google Enterprise blog post touts the application as being simple to use, with no user downtime and allowing the user to decide whether to use Notes or the Google Apps web client.

No doubt this will be of interest to those companies looking to migrate to the Google product, which combines Gmail, Calendar and other Applications, but how simple it is to use has yet to be seen, however Google believe that by using web-apps instead of a local mail server, companies can save between 60% and 75% of their budget on licensing.

If this is true, and I may be in a position to go into this further soon, then Google may well find themselves in a few more enterprise email plans than they currently occupy.

See the official Videos after the break

Steve Jobs has announced that in it’s first year the App Store has had over 1.5 Billion applications downloaded from it. “The App Store is like nothing the industry has ever seen before in both scale and quality,” said Steve Jobs, Apple’s CEO. “With 1.5 billion apps downloaded, it is going to be very hard for others to catch up.”

According to Mashable this 1.5 Billion has been hit with little over 40 Million devices having access to the system, when you take into account both the iPhone and iPod Touch. This means that on average, a user downloads 37.5 applications. To me this suggests that there are either a lot of Power users, or most people find themselves downloading the same application multiple times.

Unfortunatly there is currently no data available concerning the specifics of the 1.5 billion downloads, but I’m sure that it not only includes all the “redownloads” but also every patch too, and lets not forget when the iPhone/iPod’s latest firmware was released, most applications had to be updated to carry on working, which suggests that although 1.5 Billion is in fact a large number, perhaps it isn’t as big of a deal as we’re being led to believe.

However, that being said, it will be interesting to see how Palm and Nokia react. Their own App Stores are still in their infancy and have a long way to catch up with Apples.

In every IT Managers life, something comes along which is a bit of a challenge. Be it Political, Knowledge Based, Time Given, People Affected or simply that the task at hand is difficult no matter how much knowledge you have.

I’d like to believe that one of my current projects is all of the above, and my ego is telling me that it’s mainly to do with the fact that the task at hand is very difficult. I don’t think it is, I think it will be difficult because I’ve never done it before.

The company I work for is closing one of our sites, and migrating all of the IT down to me. This means that all systems covering that site, and the email system for a second will all have to be migrated into one. Luckily, that one system is running on VMWare, so I don’t have many of the physical issues that most other migrations face, but I do have issues of timing and complexity given the site that’s closing also runs a third sites email.

I work Automotive Industry, and we have set periods of “shutdown” where everybody except maintenance staff are forced to take holidays. This is the time when we get to change networking infrastructure, physically move equipment and play lots of Battlefield. 2 of the 3 sites involved in this migration share the same shutdown period, which would have given me a 2 week window to move everything down in a well paced, and tested way. Unfortunately the 3rd site, which is also a production based site, doesn’t. During the window I would have had to move the servers the 3rd site is still producing vehicles, which means that they need access to their email.

This then, gives me a 2 day window, which isn’t very long to move 180GB of data. Especially when Group IT policy means that if an email server is down for more than 48 hours, it doesn’t exist anymore and all mail stops being routed to it. So, 2 days, 45 accounts, 180GB of data.

The quickest way I can think of to move 180GB of data, through a 2Mbps connection with as little impact as possible is by using the compression methodolgy known in the industry as “Ford Transit”. My opposite number at Site B will be hiring a van, and physically removing, transporting and delivering the mail server as quickly as possible over the 100+ mile route on a Saturday. When it arrives, we’ll be quickly installing it, crossing our fingers, and booting it up with its original IP Addressing – only stopping to give it a new default gateway.

This should enable us to have Emails being routed again within 12 hours of disconnections. But it doesn’t help me with my server room layout, I really don’t need another physical box in that room, especially not with the ReadyNAS I’ve just ordered nor the VMWare cluster being under utilised.

So now comes the fun part, how to actually migrate everyone over, with as little downtime as possible, as soon as possible.

By clustering the Mailservers of Site A and Site B, and ensuring Site A has up-to-date databases we ensure data integrity. Lotus Domino (thus the image) makes it very simple to create a cluster, and by having 2 servers hosting the databases, they each make sure that the data on the other is up to date.

We then need to configure the cluster in such a way that instead of Site B’s server being primary for the addresses it houses, Site A’s server is. This can take some time to replicate throughout the rest of the Domino structure, and all servers throughout Europe will have to update their routings (very similar to DNS) so we’ll give it 48 hours.

Whilst the servers re-route themselves, we get everyone in Site B (and C) to set their primary mail server and Site A’s, with the backup being Site B’s. With only 45 clients, this should be achieved in about the same time as it takes the domino servers of Europe to re-route their connections. The only issue we may face with this is that for many their unread email counter will change to reflect the fact that they haven’t read all the mails on the new server because they’d read them on their old server. If anyone knows a way around this, please leave me a comment!

After 48 hours, which should be around Tuesday, we’ll be cloning Site B’s mail server into the VMWare cluster. This would technically mean that we lose a little DR, but given the cluster is pretty solid and that we’re using a very sexy ReadyNAS device (iSCSI LUNs will be used) this shouldn’t matter. It does mean however that we have a lot of data duplication, and in the future we may decide to remove Site B’s mail server completly, it will free up space, and the DR implications will be minimal because even a single server in this case will be run on a cluster. The only issue being software based, but hopefully snapshots and regular backups will aid with that.

Well, thats the plan anyway. No doubt, during the actual process things will go horribly wrong, but luckily for you I’ll be blogging about it all in filthy, dirty detail, every step of the way.

Ofcom, the UK’s Communications Watchdog, has recently announced that large areas of the country are still not able to use Mobile Broadband devices at a reasonable speed. Vast swathes of the country, and not all of them rural, have no 3G signal at all, and even where there is a signal – speeds during the day rarely get above 1mbps.

However Epitiro, a broadband communications firm, has found that off peak (at 3am) speeds of the devices alost double to 1.8mbps. This shows how massively contended 3G devices are during the day, and it’s set to get worse.

Between February 2008 and February 2009, an extra 2 million Mobile Broadband devices were sold, yet during the same time period there was next to no investment in the infrastructure providing this service.

Large areas of Scotland, Northern Ireland, Wales, Northern England and The South West are completly without 3G coverage, and yet 60 out of 65 high street stores we contacted in these 3G NotSpots were still selling the devices, and selling them as Mobile-Broadband.

Consumer groups have lept on the report from Ofcom and are blaming both the industry and the Watchdog for not doing enough to ensure that consumers are getting what’s promised.

It’s been a long time since I last posted on this blog. So long in fact that I thought it would be best to simply start again from scratch. Bluenoseddog.co.uk has been around for quite a long time, and has never been as successful as when it was  a technology blog. During it’s peak, we’d be serving tens of thousands of visitors a month, and that’s something I’d like to get back and build up as quickly as possible.

However, there will be some changes to the last incarnation of Bluenoseddog.co.uk.

There will be much less reliance on reposting news found on other sites. Instead, Bluenosddog.co.uk will be a mixture of my views, reviews, how-to’s and overall take on the technology industry as it stands today. This will mean that there will be a lot fewer posts than there were before, but hopefully it will mean that they are a lot more useful as they will be unique to this site and obviosuly my personal views.

I’ll also be trying to do a lot more of the How-To’s I started with the last blog. They were by far the most viewed posts, and given my reliance on several technologies to do my job, I think it would be a pretty simple task to put together a whole range of videos.

However, as always, I would appreciate as much feedback as possible, and if you feel you have something to contribute to the site please get in contact.

I have a few things to sort out for now with the setup of the site, but hopefully we’ll be up to full speed very very soon.