|
H110Hawk posted:If you have your entire environment on a single failure domain there will come a day when N-nines won't save you. If that is an acceptable business risk that is fine, but it should be documented in a way that upper management understands and signs off on. Entropy is a cruel mistress. You personally cannot manage literally every aspect of everything. The more humans involved, the higher the chance something goes wrong. This is why runbooks exist, because they are not human and thus their memory doesn't fade or fatigue. Following your exact same argument, the power supply in the standalone unit could fail at the same time your entire vsphere deployment does. There is no such thing as perfect security, and having a separate hardware sever does nothing to improve your chances over putting that same sever in the cluster.
|
#
?
Mar 13, 2016 17:28
|
|
|
|
| # ? Jun 6, 2024 16:28 |
|
|
Thanks Ants posted:If the display is cracked but the phone still works then I assume they just want to make sure the data is backed up so they buy a new phone and can pick up where they left off. Kinda agreeing here. On top of that, suggest the new phone has accidental damage warranty for extra points.
|
|
#
?
Mar 13, 2016 17:30
|
|
|
Sistergodiva posted:So, I have a interview tomorrow. It's for a support position. Your answer depends on the tools your employer has equipped you with for success. How would you, as a customer, expect the issue to be rectified? Is this a position at a cell phone store? Is it an IT position at a company whose IT department handles corporate mobile devices? Is it a MSP that contractually supports devices but does not sell them? All these types of employers' answers would be different.
|
|
#
?
Mar 13, 2016 17:45
|
|
|
KillHour posted:Following your exact same argument, the power supply in the standalone unit could fail at the same time your entire vsphere deployment does. There is no such thing as perfect security, and having a separate hardware sever does nothing to improve your chances over putting that same sever in the cluster. You are correct in spirit, however your example is wrong. Following my logic the building power could fail, as it is the example shared failure domain. Sure the backup physical server could poo poo the bed at the same time as the "vsphere" on a single 100% internally redundant SAN but it at least is not within the same failure domain for a cascade failure. Also why limit it to just the power supply? It's much higher odds the server dies if you include anything in the server which causes it to crash. Nothing is a magic bullet which grants you 100% uptime, and I've never stated as such. Everything is risk management. I've also stated generalized philosophies while you are sticking to specific examples. All of this being said, our environment has several single points of failure which management understands the monetary trade off of and has signed off on as an accepted risk. For specific examples: There are some systems which I cannot name here but which only exist in a single site. There are a few ways to deal with it, but one involves millions of dollars in spare hardware kicking around doing basically nothing, and the other involves THE CLOUD, a huge S3 bill, and a shitload of testing and development to even make it a viable backup solution. Our HA netscalers are another example of a single failure domain handling an entire sites data flow.
|
|
#
?
Mar 13, 2016 17:46
|
|
|
I cracked a Nexus 4 bought from Google. Called their support and even though they could do nothing to help, I would still leave them a positive feedback. The guy was very sympathetic, said he's done the same thing and it sucks, played the "let me go to bat for you with my supervisor to see if I can get this replaced under some other rule", and ultimately gave me the support number for LG to get the screen replaced. Nice and professional the whole time.
|
|
#
?
Mar 13, 2016 17:47
|
|
|
Pretty much just express empathy for the customer as we have most likely all been in a situation where we brick a phone. There's the troubleshooting part yea but I imagine this is the question to determine how good at client facing you are
|
|
#
?
Mar 13, 2016 17:52
|
|
|
H110Hawk posted:You are correct in spirit, however your example is wrong. Following my logic the building power could fail, as it is the example shared failure domain. Because two servers in a vsphere cluster are NOT in the same failure domain. There is no single thing that will bring down both of them that wouldn't bring down two standalone servers.
|
|
#
?
Mar 13, 2016 17:56
|
|
|
Nerdrock posted:Your answer depends on the tools your employer has equipped you with for success. This is for a company that Apple outsources their support to. So basically apple support since I don't think they have their own support.
|
|
#
?
Mar 13, 2016 18:05
|
|
|
Sistergodiva posted:This is for a company that Apple outsources their support to. So basically apple support since I don't think they have their own support. In that case, your employer will have equipped you for a script to handle the situation. Add a dash of personal empathy, and you're done.
|
|
#
?
Mar 13, 2016 18:09
|
|
|
Sistergodiva posted:This is for a company that Apple outsources their support to. So basically apple support since I don't think they have their own support. They do, but you have to live in Sacramento CA
|
|
#
?
Mar 13, 2016 19:14
|
|
|
Here's what you do. Buy an iPhone. Smash the screen. Call support. Take good notes on what they say, and repeat to the interviewer.
|
|
#
?
Mar 13, 2016 19:47
|
|
|
stubblyhead posted:Here's what you do. Buy an iPhone. Smash the screen. Call support. Take good notes on what they say, and repeat to the interviewer. Pretending to have an iphone thats smashed is a cheaper solution with the same result. Less amusing for the thread though...
|
|
#
?
Mar 13, 2016 21:39
|
|
|
LochNessMonster posted:Pretending to have an iphone thats smashed is a cheaper solution with the same result. Look do you want the job or not
|
|
#
?
Mar 13, 2016 21:55
|
|
|
Guys, I think this might be what they're looking for. https://www.apple.com/support/iphone/repair/screen-damage/
|
|
#
?
Mar 13, 2016 22:46
|
|
|
KillHour posted:Guys, I think this might be what they're looking for. Thanks! And thanks everyone else! 
|
|
#
?
Mar 13, 2016 23:22
|
|
|
LochNessMonster posted:Pretending to have an iphone thats smashed is a cheaper solution with the same result. All I'm seeing here is someone who isn't fully committed.
|
|
#
?
Mar 14, 2016 00:35
|
|
|
H110Hawk posted:You have simply not worked with SAN's long enough. Give it time, HA mode will wedge, a second path will be silently down, etc. Everything dies. Fair enough. I'm not a SAN guy, I've done some brocade work and the SAN guys are always trying to push bullshit iSCSI onto my networks, but I honestly can't remember a time in 20 years that I've seen an enterprise SAN array go completely dead. Evidently I've been blessed by the gods in this.
|
|
#
?
Mar 14, 2016 06:23
|
|
|
DigitalMocking posted:Fair enough. I'm not a SAN guy, I've done some brocade work and the SAN guys are always trying to push bullshit iSCSI onto my networks, but I honestly can't remember a time in 20 years that I've seen an enterprise SAN array go completely dead. Evidently I've been blessed by the gods in this. What's wrong with iSCSI?
|
|
#
?
Mar 14, 2016 06:39
|
|
|
Methanar posted:What's wrong with iSCSI? I don't like it. I've never liked it. I remember sitting in a really pretty auditorium in late '98 or early '99 in the Silicon Valley while a smarmy team of Cisco and IBM sales guys told us all how iSCSI was going to revolutionize the world, super cheap, "just as fast" as FC storage, far superior to SAN fabrics in every way. FEH. What followed was 4 years of complete poo poo as people rushed into it without understanding the real impact on the network, the performance was balls awful, switches didn't have the horsepower to handle it, the first generation IBM iSCSI network cards would loving melt themselves handling the traffic. Lets not even talk about how badly most switches handled the transition to Jumbo Frame support specifically for iSCSI. Switches today have plenty of power, its a mature, stable platform with a very low barrier of entry for small and medium businesses and I'll hate it until the day I die.  edit: oh yeah, lets also not forget the complete shitshow that was the "microsoft storage server"
|
|
#
?
Mar 14, 2016 11:56
|
|
|
So you're holding on to ideas you had about 1.0 technology from 18 years ago?
|
|
#
?
Mar 14, 2016 13:03
|
|
|
Holy poo poo 1998 was 18 years ago 
|
|
#
?
Mar 14, 2016 13:08
|
|
|
Agrikk posted:While we are here, what are your superstitions that you know are superstitions but you do them anyways? I either hold my breath or say "shhhh go to sleep" when holding down the power button to force off a server. "Maintenance period" means "Don't make plans, and don't expect me home at any hour." Also, no-one had better say "we're done" until we're actually on our way out of the loving building. go3 posted:a physical DC is a cheap answer to very expensive problems Eloquently put.
|
|
#
?
Mar 14, 2016 13:21
|
|
|
Never ever say "this is going better than expected" or anything along those lines.
|
|
#
?
Mar 14, 2016 13:22
|
|
|
DigitalMocking posted:I don't like it. I've never liked it. I remember sitting in a really pretty auditorium in late '98 or early '99 in the Silicon Valley while a smarmy team of Cisco and IBM sales guys told us all how iSCSI was going to revolutionize the world, super cheap, "just as fast" as FC storage, far superior to SAN fabrics in every way. FEH. What followed was 4 years of complete poo poo as people rushed into it without understanding the real impact on the network, the performance was balls awful, switches didn't have the horsepower to handle it, the first generation IBM iSCSI network cards would loving melt themselves handling the traffic. Lets not even talk about how badly most switches handled the transition to Jumbo Frame support specifically for iSCSI. Still terrible. Our customers requesting iSCSI are the ones least capable of maintaining a stable network, and storage kinda needs to be reliable.
|
|
#
?
Mar 14, 2016 14:22
|
|
|
This thread has some real interesting opinions sometimes.
|
|
#
?
Mar 14, 2016 14:44
|
|
|
A $10k snag we ran into with Cisco UCS is they do not officially support fiber from their Fabric Interconnect to a Nimble SAN. They say iSCSI or bust. So to get support, they're demanding we add MDS switches or buy new controllers for our Nimble. Because so few Nimble customers ever bought fiber controllers they will not invest the time to certify support for it. Then our rep asked why we even went to fiber on the Nimble and we had to dig old emails where he recommended waiting specifically for Nimble to release it 
|
|
#
?
Mar 14, 2016 15:33
|
|
|
Judge Schnoopy posted:Then our rep asked why we even went to fiber on the Nimble and we had to dig old emails where he recommended waiting specifically for Nimble to release it Your rep recommended the system end to end? (unsupported UCS to fiber Nimble?) If so, sounds like the reps problem, not yours.
|
|
#
?
Mar 14, 2016 15:37
|
|
|
H110Hawk posted:Your rep recommended the system end to end? (unsupported UCS to fiber Nimble?) If so, sounds like the reps problem, not yours. No we had the Nimble for our HP chassis that sucks rear end and we're trying to drop. They recommended fiber from HP to Nimble. Now we want to migrate to UCS, but the fiber infrastructure already exists so we have to buy thousands of dollars of extra equipment to compensate. And our vendor forgot he recommended fiber in the first place so we were less than pleased he asked "Why didn't you go iSCSI in the first place?"
|
|
#
?
Mar 14, 2016 15:42
|
|
|
How becoming is FCoE (fibre-channel-over-ethernet) for storage?
|
|
#
?
Mar 14, 2016 15:59
|
|
|
DigitalMocking posted:I honestly can't remember a time in 20 years that I've seen an enterprise SAN array go completely dead. Evidently I've been blessed by the gods in this. About 10 years ago (holy jeebus it's been 10 years) I was working at a regional bank and had come in for my morning Ops shift early to get a jump on some things that were planned for the day. One of the Senior VPs of IT (who was a really smart and good dude who knew his poo poo) was in the datacenter working on the console of one of the servers. The Hitachi SAN was beeping about a dead drive (that it had already called in and ordered a replacement for) and it bugged him so he hit the silence button. Things indeed go silent, much more silent than they should have. The Silence Alarm button was directly next to the power button and guess which one actually got hit. Oops! It turned into a long morning and a great story to explain why there was a bit of cardboard taped to the front of the SAN covering the power button. Although not technically dead, there are ways that bad poo poo can happen to an entire SAN.
|
|
#
?
Mar 14, 2016 16:07
|
|
|
For anyone else on web development freelancing, how advantageous would be owning a mac? For moving around I only got a levono yoga and I'm debating wether to just cave in, follow the trend and just get a mac, or dual boot ubuntu or something.
|
|
#
?
Mar 14, 2016 16:09
|
|
|
Thanks Ants posted:Holy poo poo 1998 was 18 years ago gently caress I'm old.
|
|
#
?
Mar 14, 2016 16:20
|
|
|
Colonial Air Force posted:So you're holding on to ideas you had about 1.0 technology from 18 years ago?
|
|
#
?
Mar 14, 2016 16:31
|
|
|
Honest Thief posted:For anyone else on web development freelancing, how advantageous would be owning a mac? For moving around I only got a levono yoga and I'm debating wether to just cave in, follow the trend and just get a mac, or dual boot ubuntu or something. Which platform runs your favorite set of tools for work? Use that one.
|
|
#
?
Mar 14, 2016 17:00
|
|
|
Agrikk posted:While we are here, what are your superstitions that you know are superstitions but you do them anyways? Never say out loud, "Sure is quiet today."
|
|
#
?
Mar 14, 2016 17:16
|
|
|
The first thing I say every time I walk onto a site is "At least it can't get any worse!" When I leave, I say "You'll never have a problem with it again."
|
|
#
?
Mar 14, 2016 18:05
|
|
|
KillHour posted:The first thing I say every time I walk onto a site is "At least it can't get any worse!" Typical contractor always trying to drum up more business.
|
|
#
?
Mar 14, 2016 18:26
|
|
|
I start the day with "Should be quiet today, I can finally finish this easy project!" just to make sure the project will be delayed indefinitely.
|
|
#
?
Mar 14, 2016 19:13
|
|
|
Colonial Air Force posted:So you're holding on to ideas you had about 1.0 technology from 18 years ago?  Its not entirely that, one of the problems is that because the cost barrier is so low my experience with iSCSI as a freelance consultant was almost always painful. Companies that couldn't/wouldn't maintain a properly built/segregated network for iSCSI were most often the ones that implemented it. Microsoft's storage server really made the problem a LOT worse because my experience would almost always be finding these things plugged into the default VLAN serving storage and causing congestion and poor application response. Granted, I made a lot of customers happy coming in and fixing it properly (when they had managed switches... sometimes not so much  ).
DigitalMocking fucked around with this message at 19:22 on Mar 14, 2016 |
|
#
?
Mar 14, 2016 19:18
|
|
|
|
| # ? Jun 6, 2024 16:28 |
|
|
Probably having an interview in the next week or two. Kinda stumped, what's the current answer for the WHY ARE YOU LOOKING TO LEAVE question if the actual answer is, "I am way under the pay/responsibility curve and they boned me on raise/bonus even though my review was stellar."
|
|
#
?
Mar 14, 2016 19:33
|
|