Discussion in 'Business & Enterprise Computing' started by elvis, Jul 1, 2008.
well you sure showed me.....lol
still dont have a real SLA
Can I build you a network that can *tolerate* the failure of an entire switch? Sure, no worries - but you're going to lose a couple of packets and then be running with an oversubscribed device. How long ago did you *test* that Fault Tolerance too?
Yes, they have redundant PSU's and Redundant Supervisor modules, but you can still have catastrophic failures that bring down the entire system... like the case I had where a Metrologist would have been more useful than a network engineer.
Well, we do.. how much are you willing to pay?
Ask your account team about HTTS, or FTS. Don't have your CFO in the same room though, unless you have a defibrillator on hand.
Another day, another X1 Carbon off to RMA
Never buying Lenovo again. You get the two-hit combo: higher than average issues, and dealing with them for RMA is like having teeth pulled
I was thinking of getting one to replace my HP. Stay well away from them??
HTTS is great and all but standard tac has degerated to the point of not being worth engaging in most cases.
That would be my advice.
Good machines on paper, good when they work. But as mentioned they seem to have a higher than normal failure rate (particularly for the LCD panel and battery).
My pick at the moment for an ultrabook would probably be the latest XPS 13 from Dell (which is probably a strong contender to be the next standardised deployment here, although the Surface is a possibility too).
Cheers for that, What's the difference between fault tolerant and redundant?
Easy way to understand it: Do you have a RAID NAS? In that example the disk would be redundant, the array is fault tolerant.
I.e. You have more disks than you require (redundancy) as opposed to being able to tolerate a fault (disk failure).
I would contest that statement
Granted, I can only speak for the products that I supported and the ones I had visibility on (primarily Nexus and UCS)
As an example, I recently saw one of the UCS engineers take a core dump from Windows, track it back to the driver (Cisco FNIC), identify the specific function and email the developer who wrote it to tell him exactly which line of code was the issue and what he thought may be the fix.
If you have issues on these products, let me know and I'm happy to chase them up.
TAC, especially the backbone centers, have more CCIE per head than any other organisation in the world. One guy in Sydney had 5 CCIE + CCDE and is going for his sixth.
Unfortunately though, tt also depends on the issue. Most Hardware and basic configuration issues are not handled by the 'backbone' centers, but majority of those engineers are trained by backbone.
For other issues, like I said above - the Duty Manager can do amazing things. Don't hesitate to ask to speak to them. They aren't technical, but can really help on process related roadblocks, or if you feel the engineer doesn't understand the problem adequately.
Then, if you do have a particularly bad experience, fill out the survey - I can 100% assure you that they are read, by both management and the engineering leads. (and likewise if they've given you a GOOD experience! )
Those surveys are accountable to very very senior levels. If TAC do not meet the customer satisfaction target, it affects the entire company at end of year.
Honestly, my time in TAC has of course made me biased.. there are certainly engineers who's ability I questioned, but by and large, I'm very proud to have worked with most of them.
No a disk pool is still only a tolerance as its it resides on the same contoller(s), in the same racks, same power ect
Yes it has.
Obviously the example was at the most basic level speaking to the NAS itself. If I wanted to provide the topology of a HA solution across DCs I would, but that wasn't necessary IMO.
All very, very interesting. Thanks for sharing.
Still doesn't explain the 57 hours they had to wait for a critical part (not that I care any more - that company since migrated to H3C , and I quite years ago), but interesting all the same.
I'd classify this as redundant. I do heaps of deployments with 1+1 redundancy for network hardware. They'll tolerate the loss of a single switch/firewall/sbc without the loss of a packet/message/call.
Last mile to the desk is usually not redundant though.
I doubt that you will find a vendor that offers a proper SLA with penalties as a standard service offering. Many will offer a penalisable SLA only if they have a hand in the deployment and you pay through the nose for it.
On the other hand lots of vendors offer what sounds like an SLA with no enforcement behind it. That is they miss a deadline and don't need to give you squat, maybe you'll get an apology for missing it but no money/discounts etc. will change hands.
Yeah my other options are Surface Pro 3 or Dell XPS 13, or shudder, a Macbook.
Agreed if its important you should do your own sparing
We had the same problem, until we found out there was a different logging path for business and now it's OK. The other method was terrible.
Hmm, care to share? (PM if needed)
I'm fairly certain we are already using the 'correct' method, but happy to be corrected.
How you all feel about capacity planning? I just got handed the storage capacity review and i'm getting scared looking at how close we are to running overtime on our nightly syncs running into the tape backup schedules.
And thats just looking at our current data, forward looking i've been told we'll need to double our current storage in normal course of operations over 5 years and one business until will need another 10tb per year for the next 5 years, and that will go up by 5-10tb per year (around 80tb at the year mark), bloody video storage.
I'm not even sure if I should put together a document and chuck it at vendors or spend the time looking up hardware specs and sending a proposal to the cto myself. Could definitely maybe be a waste of my time.
yes but is a VSS pair 1+1 redundancy, seeing as i have seen several pairs fallover i would say no. Thats why in defense nortel SMLT is know as sometimes multi link trunk.
We do, we have different SLA categories , (response, Restoration, parts, etc) and then a service window and coverage 24x7x4 24x5xNBD etc. prices is set accordingly but all have financial penalties if the SLA is breached.