Long live legacy systems

[Admiral Huddy]

Over the last four years, the bank decided to centralise every branch to Munich. This includes certain parts of London, NY, Paris and Japan.

At London, we run the last AS/400 which is so heavily bespoked that they can't port it over without massive costs.

Over the last year, the Munich servers have been down 12 times ... compared to the AS/400 which has gone done .. let me see.. nil.. In fact, our AS/400 last went down 3 years ago and only once in the last 8 years.

Today however, something has happened that I've never come across. All the servers in Munich (except one AS/400 that is a DR machine for London) are down and in fact, they have gone to DR. ;D

According to our records, in the 20 years of using System38/AS400s we have NEVER been to DR.

One up on legacy technology!!

Here's to the AS/400 - the most reliable system in the world - EVER!!

[Daz]

I'll drink to that. One of our clients is trying to move off a heavily bespoked 400 to a different 'off the shelf' type system... which has since been heavily bespoked to replicate functionality (which is against the project brief) and had to be tied into god knows how many third party products and interfaces in order to work.

It'll all end in tears...

[Mark]

Nope, one down for a poorly designed system with no failover/redundancy. DR is for recovering from disasters, not recovering from incompetence. :)

Of course, there's no doubting that AS/400 and Mainframe systems in general are very reliable. They should be given the procurement cost. Unless that is, you get some idiot installer who botches an upgrade (hello Staffordshire Uni, I'm looking at you). Mind you, god help you if they ever do break. :)

[Admiral Huddy]

haha check this out :



--> Die K-Fall-Umgebung ist momentan ebenfalls instabil.
The Disaster Environment is currently also instable


--> die bis gestern in Betrieb befindliche Produktionsumgebung von SUMMIT wurde mittlerweile mit geänderten Versionen der Datenbanksoftware Sybase ausgestattet.
The up to yesterday’s production environment from Summit was meanwhile modified with amended versions of the Database Software Sybase



Eine verläßliche Aussage darüber, ob mit der neuen Version Verbesserungen / Behebungen erzielbar sind, kann nur durch einen umfangreichen Massentest erreicht werden.
A reliable statement about whether with the new version improvements / fixes could be achieved can only be given after a comprehensive mass test



Daher wurde folgendes Vorgehen entschieden:
Therefore following procedure was decided:



1. Die jetzige K-Fall-Umgebung wird sofort gestoppt und steht damit für Eingaben nicht zur Verfügung.

Bitte Geschäftseingaben ab sofort stoppen

The current Desaster Environment will be stopped immediately and is therefore not available for any entries.




so in a nutshell, the DR is also out. A global bank currentl has no dealing system in which to trade ...

I'm off now to the job center..

;D

[Garp]

That is really crap DR, and the lack of HA on something so absolutely critical to a company is moronic :) There are a few systems kicking around work with thousand plus day uptimes on them, all in reasonably heavy use.

[Admiral Huddy]

The things is we do a test once evey 6 months and so do HO iirc. The problem is that it's not a real test. I'll get a phone call and they ask me to check the live system for codes etc.. plus they have a couple of days... When the batch results match, they call it a result. #nods head#

I have asked for a "dress rehersal DR", whereby a DR test comes unannounced and at any time but three things prevent this 1. HR due to limitation of maximum working hours, 2. Health and Saftey and 3. Cost..

F** wits!

[Mark]

I suspect that will cost them money, and lots of it. Personally, I think that would warrant someone senior getting fired, but I suspect it'll be a minion getting the blame, and given German employment law, even a disciplinary would surprise me.

HR and Health and Safety? What the hell do those have to do with testing DR procedures? In a real disaster do they seriously intend asking HR if it's OK to employ the staff and check with H&S that no-one is going to get hurt pushing a few buttons. Dunno about Northern Rock - with attitudes like that I'd be worried about having any money there.

I hope they have a proper review of what happened, why, and how to make sure it absolutely never happens again.

[Garp]

When we had one of our old filers go tits up one of our guys pulled 24 hours straight working on it before the boss twigged and got someone else to take over. Sometimes the brown stuff hits the fan and you know when you take on the job that this is always a danger and there are always going to be chances you'll have to pull out all the stops.

[Admiral Huddy]

HR and Health and Safety? What the hell do those have to do with testing DR procedures?

I know - stupid or what.. Basically, the HR element is people working more hours than the EU union allow. Not everyone has signed disclaimers. If we have people working round the clock all hours then they are entitled to stop work when their hours are up. In a non-DR situation, then there won't be the cover for the remainder.

As for the H&S it's having people on site when the souldn't be. This includes on and off site.

As i said, it's all bollox!

[Mark]

Yes, it is. May as well not bother with a DR system if that's the attitude. Of course, I can understand them not wanting workers working out-of-hours in a non-emergency situation but that's no excuse to avoid testing the systems.

[kaiowas]

A global bank currentl has no dealing syste in which to trade ...

Am I right in thinking that if anything dramatic were to happen to the markets that this bank would lose absolute bucketfulls of cash?

[Admiral Huddy]

I think that's already happening as we speak.. However, the trades will be backvalued once their in but they need somewhere to go. I'm not sure what's going on at the moment.. Another late night..