|
Tab8715 posted:That's best part about corporate IT. that sounds spooky as gently caress
|
# ? Nov 24, 2014 21:35 |
|
|
# ? May 12, 2024 08:36 |
|
FISHMANPET posted:Interview in 1 hour, bundle of nerves, etc etc. Good luck! Did mine, wasn't bad. not needing the job sure helped with not having the usual level of nerves. The rest indicator of how it went will be if they make an offer or not though.
|
# ? Nov 24, 2014 21:37 |
|
Thanks .net updates, 3 hour windows update sessions for an OOTB 2012R2 server to current.
|
# ? Nov 24, 2014 21:38 |
|
Docjowles posted:that sounds spooky as gently caress Best post in the thread.
|
# ? Nov 24, 2014 21:49 |
|
ghostinmyshell posted:Thanks .net updates, 3 hour windows update sessions for an OOTB 2012R2 server to current. And this is why I keep VMware templates of 2008 R2, 2012, and 2012 R2 patched and current.
|
# ? Nov 24, 2014 22:26 |
|
Tab8715 posted:That's best part about corporate IT. Happened to me too. Sat on an install, had done my part of the job. Watched a season of supernatural while the data guy upstairs was panicking and working like crazy. I didn't want to head back to the hotel as that mean I would have to pay for wifi and stop getting paid.
|
# ? Nov 24, 2014 23:15 |
|
Richard Noggin posted:And this is why I keep VMware templates of 2008 R2, 2012, and 2012 R2 patched and current.
|
# ? Nov 24, 2014 23:18 |
|
Docjowles posted:that sounds spooky as gently caress Load up Doom on one of the company projectors
|
# ? Nov 24, 2014 23:24 |
|
Docjowles posted:that sounds spooky as gently caress I see what you did there
|
# ? Nov 24, 2014 23:34 |
|
Misogynist posted:Briantist has some PowerShell scripts to automate this, he should probably post them. I'll elaborate a bit - we keep "template" VMs running, with Automatic Updates enabled. Periodically, we run a script that leverages ovftool to shut down the VMs, make an ovf template, and copy them to a web server. From there, we can just point VMware at the web server and we have up-to-date templates available from anywhere.
|
# ? Nov 24, 2014 23:48 |
|
Tab8715 posted:I see what you did there Today I learned Maddox still exists. Hadn't thought about that dude since about 1998.
|
# ? Nov 24, 2014 23:52 |
|
Does anyone know if any open source tools exist to scan an applications log files and look for information which may be sensitive (personal info, passwords, etc.). Ideally it should be up to our developers to have sane logging, but if a tool exists that would be handy to point these things out.
|
# ? Nov 24, 2014 23:59 |
|
FISHMANPET posted:Interview in 1 hour, bundle of nerves, etc etc. How did it go?
|
# ? Nov 25, 2014 03:14 |
|
Um, well I guess? I have no idea how to read these things though. It's kind of discouraging when they ask you a bunch of questions about technologies you have no experience with because your current employer is stuck in the early 90s. But eventually we got into the stuff I really do know (System Center), and also I learned how Senior this "Senior Sysadmin" role is. There's going to be another round of interviews with even more people next week or week after if they're still interested in me. So we'll see.
|
# ? Nov 25, 2014 05:20 |
|
JHVH-1 posted:Does anyone know if any open source tools exist to scan an applications log files and look for information which may be sensitive (personal info, passwords, etc.). Any scripting language you want. Seriously, without knowing what your logs are formatted like, this is an impossible question.
|
# ? Nov 25, 2014 06:02 |
|
Even WITH knowing, sanitization is hard as hell: IPv4 code:
code:
|
# ? Nov 25, 2014 06:10 |
|
evol262 posted:Any scripting language you want. I've seen email filters do things like detect credit card numbers so I was wondering if anything already existed. Not to detect everything, just to help find things that may be suspect. I guess I'll just take some log samplings and toss them into splunkstorm to see what the field discovery finds.
|
# ? Nov 25, 2014 06:13 |
|
Pudgygiant posted:Even WITH knowing, sanitization is hard as hell: You don't have to parse every field of every line. Email addresses alone are much harder than that article. See Mail::RFC822::Address for validating. Don't write your own validator for this poo poo, especially with regexes Use a library based on 822 for email. Try opening a socket to ipv6 (which will return an error or exception in your language of choice). Use a mapping api for addresses. You know that old jwz quip about regexes? It's not always true. But it is in some cases. Like the one I quoted. Logs have a regular format. Show a couple examples of bad lines and a competent person can generalize it. Obviously you can't catch a username or password in a log with plain regular expressions, since they're regular text and any field (auditd, kernel, system, etc) could conceivably be either. JHVH-1 posted:I've seen email filters do things like detect credit card numbers so I was wondering if anything already existed. Not to detect everything, just to help find things that may be suspect. I guess I'll just take some log samplings and toss them into splunkstorm to see what the field discovery finds. Credit cards are easily matched with regular expressions and can be determined as valid or not with simple math. Are your usernames or passwords as unusual and easy to calculate as 16-19 digit numbers with mod10? evol262 fucked around with this message at 06:25 on Nov 25, 2014 |
# ? Nov 25, 2014 06:20 |
|
mewse posted:Load up Doom on one of the company projectors Load up five nights at freddy's 2
|
# ? Nov 25, 2014 14:35 |
|
Docjowles posted:Today I learned Maddox still exists. Hadn't thought about that dude since about 1998. He still has his site but now he's added videos and somewhat of a radio thing with Dick Masterson. If you got some time to kill, it's alright. Question, I feel like whenever I read about about OpenStack I understand the product (is it even a product?) but then a week later I'm thinking to myself "What the hell is this?". Is OpenStack directly analogous to Amazon AWS (S3, EC2) and/or Microsoft Azure? As opposed as different company doing the hosting, managing underlying tech and hardware you'd be doing it all yourself. You have your own cloud platform. Am I on the right track?
|
# ? Nov 25, 2014 17:39 |
|
There's so many players involved in OpenStack that there's definitely a lot of jostling to make it "all things to all people". At least in terms of how vendors market it But yes, when you boil it down it's about running your own cloud/IaaS platform. Most of the projects that make up OpenStack map to AWS products in some way, even if it's not exactly 1:1 EC2 == nova (compute virtualization) + neutron (SDN) S3 == Swift EBS == cinder RDS == trove CloudFormation == Heat IAM == Keystone GUI Dashboard == Horizon AMI's == Glance images The same can probably be said about Azure and GCE but I'm not as familiar with those. Docjowles fucked around with this message at 18:14 on Nov 25, 2014 |
# ? Nov 25, 2014 18:06 |
|
That makes much more sense, thanks!
|
# ? Nov 25, 2014 18:21 |
|
I just moved 250 server instances from EC2 to GCE, ask me anything.
|
# ? Nov 25, 2014 19:11 |
|
Why'd you leave Amazon in the first place? How difficult was moving all the instances?
|
# ? Nov 25, 2014 19:23 |
|
Tab8715 posted:Why'd you leave Amazon in the first place? How difficult was moving all the instances? Ease-of-use falls somewhere in between EC2 and DigitalOcean. It's a commodity service, not a premium-priced offering like DO. At the same time, most resources are named, unlike on EC2 where you need to use randomly-generated instance and AMI names for everything. The CLI is beautiful and really easy to use. Because of the way GCE handles key management, it's very difficult to get locked out of an instance the way you can on EC2. It doesn't have a lot of the bells and whistles of AWS; for example, there's no RDS equivalent. I have no idea how well their load balancing system works because we use HAProxy. We're pretty much their ideal use case right now: people who are doing pretty much nothing but pure compute. Moving the instances was pretty trivial. It took us about a day to retool our Chef dependencies to abstract away the AWS bits, and a few hours to spin up new server instances. We wrote a few scripts to make bootstrapping easier -- for example, we wrote a script that locates all the filesystems on block devices attached to the box and mounts them in the right place with the right permissions before service start -- so getting the majority of servers online was basically a simple command run with a few hundred others in a for loop. Migrating might be harder if you have lots of pets and few cattle, though.
|
# ? Nov 25, 2014 20:30 |
|
Misogynist posted:Performance is very predictable because most instance types have dedicated cores, so you're not constantly running up against CPU steal issues under contention the way that you do on EC2. Since we run a real-time communications platform, predictable latency matters. Interesting, so you can't predict this enough in AWS? I'm not too well versed in cloud computing but why would this so unique or are you sharing cpu cores with other customer instances?
|
# ? Nov 25, 2014 20:54 |
|
Tab8715 posted:Interesting, so you can't predict this enough in AWS? I'm not too well versed in cloud computing but why would this so unique or are you sharing cpu cores with other customer instances? AWS runs on Xen, with all that implies. You can limit usage with cgroups, but weighted averages are kinda the norm. You're pretty much contending for CPU time on every virt platform. Some people rely on the fact that a lot of virtualized/cloud services are underutilized or memory heavy and light on CPU, so you can reliably steal CPU from other people (hypothetical 2 CPU server with 2 guests, each allocated 1 core -- if one is idling on CPU, the other one can potentially get both cores to use). That's assuming it's not oversold with 2 guests with 2 vCPUs each on a system with 2 cores, and they're both busy. It's a balancing act. And it can go the other way. So your app which expects X amount of CPU time may get less if it's on an overcommitted box or on the same machine as someone running some giant data transformation which is using 100% of their shares. The hypervisor will try to give them a little more if it's available, which can make for interesting scheduler problems. Local virt setups also have these issues, including RHEV, Hyper-V, and VMware's products, but configuring caps for hostile multi-tenant environments isn't generally a problem internally.
|
# ? Nov 25, 2014 21:06 |
|
So, GCE is doing the resource balancing act better than AWS? Or at least more reliably with less latency?
|
# ? Nov 25, 2014 21:34 |
|
New episode of Picardy Beet job extreme revenganza : did the final interview with the HR manager - even if it was biased, i was coopted by a lot of people at my future company, having worked with them before. The raise will be even more than expected (40%). And this doesn't take account of the car and mobile advantages. Really should have take the merc mentality before.
|
# ? Nov 25, 2014 21:41 |
|
Always take as much money as you can get. Your company would(and many do) fire you in an instant if they could save a few bucks a year. 'Company Loyalty' is a scam to keep timid people in their jobs for less money. A new hire at my company worked for 5 years at a large retail company with no raises. While they let go 3 employees because he worked so hard(and cutbacks). When he went back to school and was offered a new job with us, they offered him a massive raise(like 50%). He finally realized they had been screwing him the whole time. Get what you can from your job, because otherwise you will be walked on.
|
# ? Nov 25, 2014 22:06 |
|
Tab8715 posted:So, GCE is doing the resource balancing act better than AWS? Or at least more reliably with less latency? It's extremely likely that Google is heavily using cgroup cpusets, especially given how much Google uses containers internally (which were a custom cgroup thing until recently, and a lot of it still is from what I hear since not everyone's migrated from home-grown containers to kubernetes yet, and may never be, but Volmarias probably knows more). You can combine cpuset pinning and sharing so you're not sitting with idle cores, but it's a tradeoff in the same way as any sharing is, and I don't know whether their dedicated cores are actually dedicated (cpusets allocated to just your user) or just a smaller subset of the cores with a higher share ratio for your user. The resource balancing is just as configurable on AWS if Amazon wanted to do it, but it's a zero sum game. Every gain somewhere means a loss somewhere. There's no "better" or "worse", except as they relate to given workloads and provider capacity.
|
# ? Nov 25, 2014 22:16 |
|
http://www.zdnet.com/cios-make-progress-but-still-get-no-respect-7000035990/quote:The survey results showed most CEOs still regard their CIOs as itinerant specialists.… Most CEOs thought the best next step for their CIOs would be to do the same job in the same industry or in another industry. Few thought they would move on to a business leadership role. It's been reading things like this, and the general feeling I get from sitting in on some meetings with executives that make me think I shouldn't try to be the IT generalist with some business background. More and more I think I should focus on networking or SCCM or something else and just go as deep as I can in that one niche. I was talking to some old IT warhorse project managers and they mentioned that CIO's/IT directors are always the first to go when budget cuts happen or the company decides to shake things up. I guess the article isn't really a surprise, but it's still an interesting read.
|
# ? Nov 26, 2014 00:54 |
|
Don't you see though, IT only generates red numbers!
|
# ? Nov 26, 2014 01:51 |
|
BaseballPCHiker posted:http://www.zdnet.com/cios-make-progress-but-still-get-no-respect-7000035990/ Pretty interesting. I'm trying to decide between moving more towards a project management / IT manager role in the next few years or specializing in networking with a side of vmware. I definitely see this culture in our business, as our IT director is still asked to move tables and fix television sets at the chairman of the board's house. And me as the most senior technical guy is seen as nothing more than a smartphone wizard and apple tv fixer to the CEO.
|
# ? Nov 26, 2014 02:00 |
|
There are a lot of jobs where the IT department is treated really well, but they are generally going to be places where technology is integral to how the business functions and not just a convenience. Find a place where outages mean lost revenue and not merely inconvenience and you'll often find IT staff that are treated better, or at least paid better and not expected to fix the CEOs daughters laptop on their lunch break. Or go work for a technology company and become a profit center.
|
# ? Nov 26, 2014 02:28 |
|
My IT department generates revenue by reselling our internal services to outside companies who share our core business. That's a fancy way of saying other banks outsource their data processing to us.
|
# ? Nov 26, 2014 02:32 |
|
NippleFloss posted:Or go work for a technology company and become a profit center. Where people use your metrics to justify your salary--it's good and bad all at once.
|
# ? Nov 26, 2014 02:37 |
|
adorai posted:My IT department generates revenue by reselling our internal services to outside companies who share our core business. That's a fancy way of saying other banks outsource their data processing to us. My IT department generates revenue by reselling or internal services to other departments inside the company.
|
# ? Nov 26, 2014 03:05 |
|
meanieface posted:Where people use your metrics to justify your salary--it's good and bad all at once. No matter where you work someone is always trying to justify your salary based on some formula or another.
|
# ? Nov 26, 2014 03:18 |
|
|
# ? May 12, 2024 08:36 |
|
I've got some pro shortcuts for you all: 8 Game-Changing Keyboard Shortcuts You Need To Be Using
|
# ? Nov 26, 2014 05:20 |