Discussion in 'Business & Enterprise Computing' started by elvis, Jul 1, 2008.
Yes. Roast them in the survey, it gives a great sense of accomplishment.
It's not something I setup, it is something how ever I changed.
Also this site is currently setup like this for email
ISP Filter>Gateway Filter>Trend on exchange Filter>Exchange.
When I found that out I removed Trend from Exchange straight away. But as anything that that stuffs around with Exchange it wasn't a simple uninstall
In an astonishing announcement (/s) 'IBM Australia’s chief engineer has revealed a simple routine test could have avoided the collapse of Census website':
Aka... Why High device uptime is a terrible metric.
Nah, I don't need to save the running config, She'll be apples mate.
It's more a failure of test procedure and then validation of the procedure.
High uptimes are fine for systems with auto-patching and fault tolerant where systems can join/leave a cluster. I have systems over 1000 days without issues, the difference is they're all cattle not pets. I don't care about any of them individually, I only care if they're all dead at once
IBM can tick the list of horrid failures and absolutely horrid design. Their "N+1" networking is single router per single provider. They only had 512Mbit of transit via NextGen, which is why the "DDOS" didn't even register on a global scale. A few domestic DSL accounts could have saturated the connection, yet alone a few compromised websites / servers with increased bandwidth.
You NEVER do DDOS mitigation at the smallest choke point of your network, you're already at the too late stage. The design and lack of any scrutiny is a double stuff-up.
A single VM on a free trial account with Azure could have done that a thousand times over.
Agreed. It's why I always explain that just because I've saved months on the schedule does not mean we can go live early - it's for proper testing and remediation because every system has issues. As with any decent test plan, you don't have to do it if you're documenting known behaviours, but whoever signed off on that design should be fired.
they should be fired just for using the name island australia, let alone the concept.
Is it just me or has Luke212 been super quite post census night???
if you look back on my posts you would know i would have an army of whiteboxes around the country and thus be impervious to ddos. some of my ideas will catch up one day... until then you will keep getting "ibm"s
in some sort of ring I assume.
10gb token ring, using windows whiteboxes, natch.
but not 2016, because that would require powershell, and be too hard. Because developers.
The trick to a good troll is to keep it subtle.
Well they have to make it complex so they can't be replaced, le duh
2016 Storage Spaces, 2016 Hyper-V and SMB Multichannel oh my!
Is storage spaces still shit?
It wasn't ever shit. Its just geared for mirrored arrays - not parity ones.
Was moving things around at ~640MB/sec earlier.
Need my other Chelsio cards and then I can swing to the heights of RDMA, SMB Multichannel and RSS.
Oh, Hyper-V 2016 supports RDMA vSwitches (this is a very very good thing).
Parity should never have been an option, has this changed at all in 2016?