ITJuju.com

The following 2 posts were made on the warrior forum under the thread: Getresponse…your lists aren’t working…

That was before my account “savvypro” was suspended for spamming by a modaurator who seams to not have read any of my posts in that thread. And linked me in with someone who was promoting theire software as an alternitive just because I asked him a valid question which was also brought up eairler in the thread.

I posted in that thread because I’m a paying customer of Getresponse.com having problems with the service after their upgrade – just like everyone else in that thread.

Anyway I made a copy of my posts as the information in them are of importance – I was going to place an edited version of them here covering other areas as well – but now I’m just putting what I wrote to ensure the infomation is avaiable.

As it’s no longer on the forum, it’s now here.

The information by Simon Grabowski is in the gray boxes and is also italicized.

Everything else is: © MMIX ITJuju.com — All rights reserved

Originally Posted by Simon Grabowski
I appreciate your comments, however we’ve given a lot of advance notice to our users. It has been communicated via email twice, and it’s been posted in the GR panel. This is the biggest upgrade we’ve done in 10 years and there was absolutely no way to set up such a huge upgrade in a seamless way. That’s what happens when you have to migrate with 93,000+ active accounts and hundreds of millions of email addresses to brand-new platform. And it’s not just GR that does this. When Apple releases its new products, their Apple Store goes down for the entire day

I’ll preface my statement by stating that I don’t know what or how your infrastructure is like or how it’s set up, my comments are based on just the a above quote.

I have a problem with the above statement in particular the following two bits:

Originally Posted by Simon Grabowski
there was absolutely no way to set up such a huge upgrade in a seamless way.

I’ll grant you that it’s difficult to do big bang upgrades in a “seamless way” or a problem free way.

But the following is just wrong:

Originally Posted by Simon Grabowski
That’s what happens when you have to migrate with 93,000+ active accounts and hundreds of millions of email addresses to brand-new platform.

I’ve been involved in a number of upgrades which involved moving data from a number of storage servers and NAS clusters to new storage.

This included transferring data which was involved in the use of:

• 30+ active websites

• Account profiles for 30,000+ AD accounts

• The network “home” drives for the 30,000+ AD accounts

• Oracle database dumping areas

• Oracle log file storage areas (separate from the above)

• Department shares

• Application shares

• Exchange cluster

• A couple of NAS clusters

• And a whole bunch more.

If a big bang approach was used we would have ended up having to implement the disaster recovery plan which still isn’t a pleasant though even after leaving the place almost a year ago.

What we did can be sumed up with the old saying:

How do you eat an Elephant?

One bite at a time

All the data was replicated over to the new “platform” before we ever did a migration, which then meant that when it came time to move things over – there was less to copy over but even then it still took in some cases a whole night per few thousand accounts.

Everything was always tested on the IT department first as we were more likely to sport the tiny problems most users wouldn’t.

Everything would be done in batches and after each batch we paused and waited to see what had broken or who would scream – once the dust had settled the next batch was started.

OK – it did take months just to plan and even longer to implement as lots of things had to be re-coded in order to use the new SAN.

But everything worked, very little broke – when people did screamed it was always about a workaround that was used on the old system which just needed to be removed. And this only affected a small number of accounts out of the 30,000+

Although we did find that one of the senior management had given themselves access rights to all admin shares – something they shouldn’t have done. The plus side was that it broke their account and highlighted what he had done.

The new code just wasn’t going to allow him to continue breaching the data protection act.

The old hardware/systems were kept in use until everything was done and we were happy with the results. This overlap meant that we could switch back at any time – the problems (if there were any) were always limited to those that had been moved over.

By the time we got to moving the critical things the process had been tried and tested multiple time over many months.

….

As I said at the beginning – I don’t know what or how your infrastructure is like or how it’s set up, my comments are based on just the quote at the beginning of my post.

A follow up:

Again: The information by Simon Grabowski is in the gray boxes and is also italicized.

Everything else is: © MMIX ITJuju.com — All rights reserved

Originally Posted by Simon Grabowski
Thanks for your feedback. I don’t intend to argue about different
migration roadmaps because like you’ve said it yourself: each
platform is different. Transferring data of 30,000 accounts
is different from migrating 93k active users that send approx.
6 billion permission-based emails a year to an entirely new
hardware AND software infrastructure, which is a roadmap
process (and not a “big bang” approach).

Simon,

The 30,000+ accounts was just one small part of what I mentioned and was only part of just one system.

The example that I mentioned involved complete changes in both hardware and software infrastructure as it affected everything.

Yes each platform/network/infrastructure is different even when the same technologies are used.

But from where I stand as a paying customer – it looks like a big bang deployment, as your doing the change over in one go.

You may have had all the infrastructure and new code in place but the change over was in one go.

I’ve dealt with too many vendors to know what a “road map” really is.

Hopefully the minor issues that people are seeing are temporary.

If you need someone to do a few test let me know as I’m still in the process of setting up my account and am not sending signups just yet.