|
/26 and /27 are good if you have a bunch of single purpose cloud provider networks that need to be part of a wan
|
#
?
May 28, 2021 21:42
|
|
|
|
| # ? May 27, 2024 21:30 |
|
|
SyNack Sassimov posted:Also, anyone who prides themselves on having memorized the entire set is gonna have a real fun time when IPv6 really starts being used widely, which I'm sure is just around the corner. It's totally really gonna happen this time, the US federal government is handing out mandates that departments need to be 100% converted to ipv6 by the end of 2025. No dual stack allowed.
|
|
#
?
May 28, 2021 21:51
|
|
|
deedee megadoodoo posted:The correct answer to all of these dumb interview questions is google. I’ve used some version of “I don’t know off the top of my head but I can look it up when I need to.” In every drat interview for the past decade. There’s no reason to memorize trivia. goddamn, I've been giving interviews for the past week and i would be so happy to hear someone be open and honest about not knowing something a guy told me today that he liked testing because it "pissed off the devs" and i was ready to hire him on the spot
|
|
#
?
May 28, 2021 22:00
|
|
|
Defenestrategy posted:When asked about league of legends 'I dunno,, it sounds like a jank mod for warcraft' "That's the lovely version of Dota right? I might have played it once or twice. It was ok." deedee megadoodoo posted:The correct answer to all of these dumb interview questions is google. I’ve used some version of “I don’t know off the top of my head but I can look it up when I need to.” In every drat interview for the past decade. There’s no reason to memorize trivia. Sepist posted:I've been in networking for about 15 years now and /27, /28 and /29 still require mental gymnastics to convert I am counting on my fingers from the nearest octet and ain't nobody going to tell me otherwise 
|
|
#
?
May 28, 2021 22:04
|
|
|
Plus even if you memorize it, poo poo gets patched and deprecated all the time. I was getting MacOS training and the guy didn't even know Network Monitor got taken out in Catalina.
|
|
#
?
May 28, 2021 22:15
|
|
|
Cloks posted:goddamn, I've been giving interviews for the past week and i would be so happy to hear someone be open and honest about not knowing something I'm pretty sure my candor (and the interviewers being personable) in this exact area was why I was hired for my most recent job. That and my incredible CS skills, but mostly this.
|
|
#
?
May 28, 2021 22:42
|
|
|
Sickening posted:He also pushed me with question like "what does code for adding a waf in azure look like" and I could only laugh and say "exactly what the examples in Microsoft's wiki say they are" and he didn't seem amused.  I would have said it looks like an ARM Template and it's written in JSON. It contains variables, parameters and resources. Whomever actually knows the code for freaking Windows Azure Firewall off the top of their head has got to be one really freaking weird dude.
|
|
#
?
May 29, 2021 01:33
|
|
|
I love these hypothetical job situations in which you have the ability to make API calls to a public cloud provider to deploy workloads, but Oops! You can't load the docs site, Google's in their usual 8am-5pm daily downtime, and the CEO is going to personally fire you if you don't update the WAF in the next 5 minutes to block the skid bot that keeps emailing him lemonparty.jpg through the Contact Us form.
|
|
#
?
May 29, 2021 04:22
|
|
|
Cenodoxus posted:block the skid bot that keeps emailing him lemonparty.jpg through the Contact Us form. "Oh sorry sir, I'll turn that off. I was a bit late this morning so wasn't able to get in in time to prevent the dead man's switch from firing."
|
|
#
?
May 29, 2021 04:34
|
|
|
Good news, az network -h will probably get you where you need to be even if google poo poo the bed and you forget the url for the MS docs site. I have somehow managed to dodge all trivia based technical interviews so far in my career. My last interview for an Azure Cloud Engineer role involved no real technical questions outside of talking about projects I’ve worked on/architected, but nothing in stupid detail like that.
|
|
#
?
May 29, 2021 04:37
|
|
|
Sickening posted:I simply don't free type my terraform code from memory. There are some things I just wouldn't ever do from memory. I'd use a reference, either documentation, sample code, or copying what worked last time. There are common command line tools that I've used for years where one or two options are in my memory wrong for some reason and they just won't correct. God help us all if I have to use one with no references and under any pressure at all. And would you really want to deploy based on someone's memory, with no cross-checks (or testing) on the code ? Apparently that must be a thing in some shops.
|
|
#
?
May 29, 2021 14:45
|
|
|
We have a handy well written shell script that takes a hostname and a one time password to generate a key. I know the option, I know it needs the fqdn, but every single time I check the help out anyways to visually confirm because I HATE loving up the one time password and needing to ask for a reset. So I get paralyzed that I'm about to gently caress it up and double (then triple) check the arguments. I've done it hundreds of times.
|
|
#
?
May 29, 2021 14:50
|
|
|
I was in one interview about a decade ago and one of the questions was: “You have one week to build a storage platform that is infinitely scaleable and when you’ve built it, we’ll all retire rich. How would you build it?” I fumbled with the answer, coming up with various RAID arrays and data partitioning schemes but couldn’t come up with anything satisfactory. It turns out that the company was struggling with that problem internally and did not have an answer either. Using its recruitment / interview candidates as free thought labor really ticked me off then and still irks me now and I turned down their offer (which was the right call because the guy doing the hiring quit two weeks later and called me up out of the blue to say thanks for the NetApp mug and that I made a good decision to avoid the place.) The answer now of course is “Invent S3”.
|
|
#
?
May 29, 2021 16:25
|
|
|
Agrikk posted:“You have one week to build a storage platform that is infinitely scaleable and when you’ve built it, we’ll all retire rich. How would you build it?” Should have looked really thoughtful for a few minutes, looked up, and go "IVE GOT IT! I'M GONNA BE RICH" and bolted out of the interview.
|
|
#
?
May 29, 2021 16:50
|
|
|
Agrikk posted:I was in one interview about a decade ago and one of the questions was: Did.....did they ever consider putting out an RFP or even contact Storage or Solutions vendors?
|
|
#
?
May 29, 2021 17:21
|
|
|
I left a really toxic job and had to take a very minor paycut in the process, but it turns out my health insurance was so costly at my last job I'm actually getting *more* money on my paycheck. My boss appreciates me, I'm not work more than 40 hours a week, I'm learning cool stuff, and I'm getting to come up with projects to make things better. I already fixed their jacked field office printing situation. Feels good
|
|
#
?
May 29, 2021 18:28
|
|
|
Agrikk posted:“You have one week to build a storage platform that is infinitely scaleable and when you’ve built it, we’ll all retire rich. How would you build it?” "Buddy, if I had that figured out I wouldn't be sitting in this interview."
|
|
#
?
May 29, 2021 20:12
|
|
|
Okay so this is a very stupid question but in the context of routing protocols, neighbouring or adjacent routers are just routers that have a direct link established between the two of them with no hops, right? Or does some concept of “adjacency” explicitly need to be established through some form of router-router communication? Took the GCP network engineer practice exam, quickly realized I needed to know way more about routing protocols than I did. On the plus side they make a lot more intuitive sense to me now than they did a few years ago! Anyways I feel dumb even typing this because there’s no protocol specific definition and I can only assume they mean the literal definition but no resource has defined that word pairing before and it’s a rather important definition to get right. vvv this is what i meant but not what i said The Iron Rose fucked around with this message at 03:17 on May 30, 2021 |
|
#
?
May 30, 2021 01:58
|
|
Cat Army
|
I would interpret adjacent “the next hop” but am also very much not a networking person
|
|
#
?
May 30, 2021 02:33
|
|
|
The Iron Rose posted:Okay so this is a very stupid question but in the context of routing protocols, neighbouring or adjacent routers are just routers that have a direct link established between the two of them with no hops, right? Or does some concept of “adjacency” explicitly need to be established through some form of router-router communication? I never thought "next hop" was actually official. I just took it to mean each stop for a return in tracert or similar path detection. I would also assume the routers links are layer 3 and this should normally fit criteria as its going to show up in the trace. If the links are more like your stacking cable or some similar technology, it might not apply.
|
|
#
?
May 30, 2021 02:40
|
|
|
Also, GCP networking is the worst because of their SDN overlay. There is a ton of limitations. Private GKE cluster? Sorry you can't hit the master unless you use a bastion. Want to route traffic between two instances in the same network? Sorry all traffic must go to the subnet default gateway first even if you configure the instances to use the real subnet mask (all GCP instances are a /32 by default if I recall). Using VPC peering and advertising a 10/8 code? Sorry now you can't create GKE clusters because the overlay thinks you're using all available subnets. Just maddening. Sepist fucked around with this message at 12:50 on May 30, 2021 |
|
#
?
May 30, 2021 03:06
|
|
|
The Iron Rose posted:Okay so this is a very stupid question but in the context of routing protocols, neighbouring or adjacent routers are just routers that have a direct link established between the two of them with no hops, right? Or does some concept of “adjacency” explicitly need to be established through some form of router-router communication? A direct link between routers is default behaviour but not always required. For example, OSPF virtual links and BGP multihop allow you to form adjacency/peering between routers not directly connected. As for discovering neighbours, OSPF will send out hello packets automatically to discover them, BGP you will need to explicitly configure which AS's and neighbours it will peer with.
|
|
#
?
May 30, 2021 10:41
|
|
|
The Iron Rose posted:Okay so this is a very stupid question but in the context of routing protocols, neighbouring or adjacent routers are just routers that have a direct link established between the two of them with no hops, right? Or does some concept of “adjacency” explicitly need to be established through some form of router-router communication? I haven't had to think about this for a while, but IIRC adjacency is a specific relationship, not just ones that are connected to each other, although they will try to form an adjacency in that case. There are specific requirements for adjacency, and they aren't the same between routing protocols. In OSPF, for example, routers will not form an adjacency if they are configured in different OSPF areas. Mismatched hello and dead timers will also prevent adjacency. I would have to go dig out my old CCNA books to get all the details, but if you aren't able to find them, I will do so.
|
|
#
?
May 30, 2021 10:42
|
|
|
The variety of responses here reassures me that it was not a stupid question, as clearly it’s an under-defined concept! But virtual links and of course BGP peering make sense now that I think about it, and the detail about OSPF zones was key. Thank you all so much!
|
|
#
?
May 30, 2021 17:12
|
|
|
Adjacency implies routing information is being exchanged. This implies some sort of session creation (TCP 179 for BGP, IP proto 89 for OSPF, etc.) There's also a discovery prerequisite for this session setup. Multicast discovery is a common approach , static endpoint definition is the other.
|
|
#
?
May 30, 2021 17:54
|
|
|
Are flat ethernet cables bad? Seems like they'd be more susceptible to interference but google seems to be divided on the issue
|
|
#
?
Jun 1, 2021 14:56
|
|
|
Ethernet cable design relies on the twisted pairs to protect from interference from other wires in the cable. So a flat cable can probably work, but only over very short distances. I've never bothered to try and we don't keep flat cables in house anymore so the opportunity is gone.
|
|
#
?
Jun 1, 2021 15:24
|
|
|
I have a load of flat cables that came with our Cisco video systems. They all seem to work great.
|
|
#
?
Jun 1, 2021 15:29
|
|
|
A new (female!) engineer that I interviewed just started yesterday, will be joining my team in a month after she finishes onboarding. I just heard from her that she negotiated with three different people to make sure that she was going to be working with me specifically, and that she was super excited to start. Just such an incredibly rewarding thing to hear.
|
|
#
?
Jun 1, 2021 15:40
|
|
|
Hell yeah! That's awesome.
|
|
#
?
Jun 1, 2021 15:52
|
|
|
xzzy posted:Ethernet cable design relies on the twisted pairs to protect from interference from other wires in the cable. So a flat cable can probably work, but only over very short distances. I've never bothered to try and we don't keep flat cables in house anymore so the opportunity is gone. right, I know they're usually twisted I just wondered if these had a different kind of shielding or something. I see pictures of people using them in racks and it looks so pretty and I would love to swap out about 300 of these gigantic shielded cat6a cables I have across the IDFs here (idk why they even used 6a, none of the switches in production can even do better than a gbit over ethernet anyway)
|
|
#
?
Jun 1, 2021 16:05
|
|
|
I had a bunch of flat cables that ended up being bad. They were all from the same order/batch, so I'm sure that had something to do with it, but it was enough for me to rip them all out and replace them and not order flat cables again.
|
|
#
?
Jun 1, 2021 16:14
|
|
|
The Iron Rose posted:A new (female!) engineer that I interviewed just started yesterday, will be joining my team in a month after she finishes onboarding. I just heard from her that she negotiated with three different people to make sure that she was going to be working with me specifically, and that she was super excited to start. That’s really cool, but how does onboarding take a month?
|
|
#
?
Jun 1, 2021 16:16
|
|
|
I assume that includes notice, probably a week or two between jobs, etc.
|
|
#
?
Jun 1, 2021 16:18
|
|
|
Last place I worked it could take a week or more to get an email address if you previously worked anywhere else in the organization -shrug-
|
|
#
?
Jun 1, 2021 16:40
|
|
|
The Fool posted:That’s really cool, but how does onboarding take a month? Every new hire, including C levels, works in customer support answering phone calls and emails from customers so they learn the product. It’s actually really cool, it’s good for people to get a dose of humility and compassion for our support folks. So she’s an employee as of today, but won’t start tech work till July. There’s some basic training and access provisioning that happens in that third week, but for the most part it’s hands off for that first month. E: you get paid your regular salary during this time too ofc. You also need to “volunteer” to work on at least one public holiday so the full time support people don’t have to. Which is not that bad, I answered emails while relaxing in the sun in Mexico (that’s a lie I spent it crammed in the only room in the villa with Internet going through nicotine withdrawal, but in theory lmao) The Iron Rose fucked around with this message at 16:47 on Jun 1, 2021 |
|
#
?
Jun 1, 2021 16:42
|
|
|
thats loving ridiculous - who allowed such a clearly sensible and valuable policy to be implemented and followed??
|
|
#
?
Jun 1, 2021 16:43
|
|
|
The Fool posted:That’s really cool, but how does onboarding take a month? Some orgs have extensive orientations. A month sounds super long but I am also not shocked. The Iron Rose posted:Every new hire, including C levels, works in customer support answering phone calls and emails from customers so they learn the product. You definitely are getting a certain type of personality when hiring with that kind of stipulation. I would think a large part of the talented workforce would balk at that kind of requirement as it sets some red flags. If this was shopped to me as part of the interview process, I would label you as a company trying to get me to drink the cool aid and nope out. Sickening fucked around with this message at 16:56 on Jun 1, 2021 |
|
#
?
Jun 1, 2021 16:49
|
|
|
That's pretty cool, but such a foreign idea to me. The new place has a "week" (sounds like really two days) orientation and I almost don't know what I'm going to do with myself. I'm so used to being dropped in the fire.
|
|
#
?
Jun 1, 2021 16:51
|
|
|
|
| # ? May 27, 2024 21:30 |
|
|
It could also be a great way to highlight ridiculous workflows that support staff are subjected to but nobody cares to fix because "it's only support, they can deal with it", whereas a C-level coming on board has the ability to make changes or at least give an outsider's perspective.
|
|
#
?
Jun 1, 2021 17:07
|
|
