Discussion in 'Business & Enterprise Computing' started by elvis, Jul 1, 2008.
can't believe they didn't go for IDDIOT, the Intel DDIO Threat
Not in mega cloudy land, but face the same issues at a smaller scale. Either way, there's simply no viable option to throw humans at the issues I face and automation is the only way. And, it works. Have smart employees and smart systems and simply get shit done.
I use the 10x rule. Dealing with 20 or 30 servers and hand updating them as pets is workable. Hand updating 300 servers isn't (I wish I only had 300!), you're now dealing with cattle. Plan all changes as if you're dealing with 10x what you currently have. If you set your target high enough, you'll implement systems to cope and as companies and systems grow in size. If you stick to your 20 server mindset, you'll only ever cope with this number.
80% of my role is to automate and orchestrate. If I'm doing something manually more than once, it needs to be automated. Even this week I replaced a mandraulic task which was only taking 2 hours a week a year ago to do manually but now with growth it's 4-5 hours a week worth. It took about 30 hours of development time to complete and will work with 10x the systems without any changes. Next week we'll pick a similar task and do the same thing, some are 100+ hours to complete so they're not always quick wins but each time the systems do the work not humans.
The biggest issue is that corporate and MSP worlds love to just throw the wrong sort of people at the problem. They throw lower end workers, not high-end engineers. This used to be part of the Linux vs Windows sysadmin disparity in terms of numbers, Linux sysadmin work used to be far more complex and therefore only those who were capable of it worked on that side of the fence. Greater comprehension leads to smarter decisions, which leads to more systematic approaches instead of benign data entry approaches older Wintel sysadmins had. Tables have turned now and I think Linux is actually easier now that people are used to containers and orchestration, but there's still millions of environments still stuck in the dark ages of IT.
That was my thinking at my last job, went from 3 VMs at HQ and about 7 "servers" at remote sites (8 sites total) to about 18VMs at HO/admin sites and 6 sites with 2x VMs per site (12 sites total with 3 more coming on). HQ was mostly auto once SBS got murdered in the night.
Remote sites needed to be treated as pets due to POS vendors idea of patching being "team viewer on, stop services, copy over new files, start services, bail" - I was hoping to get a dump of the updates from them to HQ then I couldve auto deployed it overnight at 3am to be safe for testing in the morning by venues :/
Auto tasks like reboots/certain pos related service restarts overnight/clearing queues and caches/syncing and exporting etc were all auto. Had a giant folder of scripts/ideas in-progress for removing all my headaches.
MSP share holders dont get rich if the customers only spend up once a year, break+fix is the key to lasting business
Was on call with a bloke for 2 hours because his network is trash. Spent 30minutes of the call explaining what a timeout was and why 5 minutes to get a connection over telnet isn't optimal.
Liar! Obviously the root cause is your telnet server doesn't have enough CPU time available. Everyone knows telnet needs 8 CPU cores, 32GB of RAM and cannot run virtualised! This is all your fault!
Bet the problem is when he replaced one of the hard drives with a non-vendor approved hard drive. If you had just paid $3499 excluding GST for the vendor approved 500GB SATA drive you wouldn't be in this hot mess. Remember, vendors know best.
When it comes to vendors being assholes with hardware prices, industrial machines with a brand new WinXP machine attached to them take the cake. "Mmmm the PC for your 15 million dollar steel cutter is broken! we don't have parts despite it being fundamentally standard PC hardware that we have intentionally made proprietary through the use of QC stickers and security screws. Your only option is a new PC running the latest Windows Vista and this costs $60,000 because we have to get it in from germany, hand crafted by AMG engineers when they are not busy making ze automobielz"
and because it's running a 15m machine, that's costing the customer bulk coin being down, they'll happily throw the money at the vendor.
Monday... so you need help with IT on something totally unrelated to the project I'm doing for you and its URGENT... Sure is, just like your 14 day late payment.
More ransomware insurance mobs popping up:
What percentage do you think just negotiate with the ransome dudes a lower price and then just pay it ?
I'd guess, a fairly high % !!
There was a well know news article a few months ago, where a company saying it could decrypt your files (for a fee of course), was found out to simply negotiate a lower price, pay off the bad guys pocket a profit, end user got warm fuzzies for paying less than the ransom (and not rewarding the bad guys - if only they knew), and company X got to profit from cybercrime as well as the baddies. win win right?
As the register article postulates, this will only serve to increase attacks and the requested ransom amount. and who wins? the lawyers insurance companies and the scammers. You can bet the currently low premiums won't last, as they start paying out more and more ransoms.
Maybe the insurance companies will start stipulating minimum requirements to lock down and backup your shit as part of the insurance deal, so you don't get crypto'd in the first place. Something that an IT department can take to the CxO's to get the resources that they've been asking and being denied in the past. That's the only good I can see coming from it.
This always comes to mind whenever someone brings up cyber insurance. If you get hacked most of the time it’s due to poor security practices. Proven poor security = can not claim.
It’s like claiming insurance after you have intentionally burned down your car.
security is expensive and 'poor' is a relative term.
some people would not know its bad to leave a fuel canister in the trunk on a hot day.
Christ we didn't even make Monday midday without a car analogy.
I didn’t know about that. If anyone asks I still don’t. “But there were no warning labels or nothing!”
The assessor might scratch their head if you have a can of petrol in your diesel truck though so use common sense
you in straya mate.
True story. A friend bought their first car, and it was a bit of a beater. His old man, being a bit of a car guy, called and got insurance that morning for it, and then went out to the car to fix one of the seat brackets and welded it up. Went off to do something else and the weld caught the interior on fire and the car burnt out. Called insurance back that afternoon and put in a claim, surprisingly went though after a lot of questions asked.
No utes left with Ford and Holden no longer making them.
All we're left with a girly trucks made in foreign countries.
And now with HSV and Dodge importing Silverado/RAM we have 'murica freedom edition* trucks now too.
*i.e. 4x bigger than they need to be
trunk not truck