|
madsushi posted:> wrote a lot of OSX software To be fair, while OSX sucks at a lot of tasks (and especially a lot of server tasks, and realtime tasks that touch the kernel, etc), Quartz is really good, and it's not trivially portable. If they wanna throw money away, their call
|
#
?
May 7, 2015 02:43
|
|
|
|
| # ? Apr 30, 2024 18:27 |
|
|
madsushi posted:> wrote a lot of OSX software This is pretty much my takeaway. "Welp our app has to run on OS X for $stupid_reasons. Guess we have no choice but to waste millions of dollars on insane bespoke racks and machines that do half the work at twice the cost of a normal loving blade server". The is the most apple.txt thing, and I say this as a big Apple fanboy. Would like to be a fly on the wall at the board meeting where they justify that expense.
|
|
#
?
May 7, 2015 02:44
|
|
|
See, this is why I like consulting instead of operations. If I recommend a customer fix something that's obviously stupid and they choose not to, they have to live with it every day, not me.
|
|
#
?
May 7, 2015 02:48
|
|
|
If anyone cares that much, the article is up on Hacker News. Poster "skuhn" works for the company (and in fact wrote the original blog post) and has thrown up a shitload of responses about their reasoning for going with the iTrashcans.
|
|
#
?
May 7, 2015 04:02
|
|
|
evol262 posted:To be fair, while OSX sucks at a lot of tasks (and especially a lot of server tasks, and realtime tasks that touch the kernel, etc), Quartz is really good, and it's not trivially portable. If they wanna throw money away, their call It just seems strange to write a web service for something that locks you into Apple hardware.
|
|
#
?
May 7, 2015 04:09
|
|
|
Inspector_666 posted:It just seems strange to write a web service for something that locks you into Apple hardware. Honestly, write whatever's cheapest and gets your product to market fastest. You can always rewrite later, and it's not like a dumb image processing pipeline is some multimillion-dollar rewrite. Docjowles posted:This is pretty much my takeaway. "Welp our app has to run on OS X for $stupid_reasons. Guess we have no choice but to waste millions of dollars on insane bespoke racks and machines that do half the work at twice the cost of a normal loving blade server". The is the most apple.txt thing, and I say this as a big Apple fanboy. Given their image quality targets, and the fact that Core Image is GPU-accelerated and can produce really great quality images with very minimal CPU usage, it's entirely possible that they were able to deliver significant cost savings at scale by doing this. Vulture Culture fucked around with this message at 04:52 on May 7, 2015 |
|
#
?
May 7, 2015 04:39
|
|
|
Docjowles posted:If anyone cares that much, the article is up on Hacker News. Poster "skuhn" works for the company (and in fact wrote the original blog post) and has thrown up a shitload of responses about their reasoning for going with the iTrashcans. From what he's saying it kinda makes sense. It just seems like the stupidest thing ever out right, but given that Racklive did the design and build of the holders (so they didn't have to care about that part) and that this has to last just for a depreciation cycle (2-3 years), you can get some good value and money out of it for the time being, especially if https://news.ycombinator.com/item?id=9502091 posted:Here's the quick math on cost per gflop, including all network and datacenter costs: You can get a lot further ahead of your competition if their costs are 4x yours (he mentions later down that all their competitors are in EC2).
|
|
#
?
May 7, 2015 04:52
|
|
|
Will scaring the hell out of someone by yelling "NO WAIT DON'T" when they say they're rebooting something, ever stop being funny? My sources say no.
|
|
#
?
May 7, 2015 07:47
|
|
|
MC Fruit Stripe posted:Will scaring the hell out of someone by yelling "NO WAIT DON'T" when they say they're rebooting something, ever stop being funny? My sources say no. That old chestnut is right up there with saying "Rebooting (production) server now" when the dev guy asks to bounce a similarly-named QA box. Panicked looks are like gold, man...
|
|
#
?
May 7, 2015 09:25
|
|
|
Anyone here (or have ever been) a Help Desk/Tech Support Manager? I just figure that to be a sort of missing link between Network Admin and Helldesk
|
|
#
?
May 7, 2015 14:58
|
|
|
Inspector_666 posted:So this company called imgix is very proud of their rackmounted Mac Pros. This is hilarious and I can tell that the guys who run this company are ~~~ MAC GUYS 4 LYFE ~~~~~ just by the design of the website. Seriously does OSX even have clustering capability these days?
|
|
#
?
May 7, 2015 15:07
|
|
|
have we covered that this new design of one extremely long web page is loving horrible
|
|
#
?
May 7, 2015 15:08
|
|
|
go3 posted:have we covered that this new design of one extremely long web page is loving horrible http://www.eskaustin.com/ I think this is the worst one I've found yet. I cannot figure out a drat thing from this website. I just want to see what the gently caress you're about and how delicious your food is.
|
|
#
?
May 7, 2015 15:14
|
|
|
Race Realists posted:Anyone here (or have ever been) a Help Desk/Tech Support Manager? I just figure that to be a sort of missing link between Network Admin and Helldesk Ehh...not really. The Manager bit means that while you're still on the hook for any sufficiently large technical issue (though that varies with environment/company) your job is primarily herding cats. Depending on the competency of your helpdesk personnel your job will be easy or a living goddamned hell where every other manager+ in the company who has a service complaint rolls downhill into your front door. A more bridge-y position is Senior/Level 3/whatever Tech Support, then if you're friendly with the network guys you can generally sit them down and ask if they need a Jr Admin. It can be difficult to make the shift, to be honest--helpdesk is end-user support and lap/desktops, network is servers, infrastructure, and other back-end bits. Depending on the environment, helpdesk is expected to solve the problem NOW, whereas unless its actively on fire most network stuff is supposed to be done in a more methodical manner during maintenance windows, possibly with a fleshed out rollback plan.
|
|
#
?
May 7, 2015 15:31
|
|
|
Thanks for the answer. back to learning PowerShell I guess
BornAPoorBlkChild fucked around with this message at 15:43 on May 7, 2015 |
|
#
?
May 7, 2015 15:39
|
|
|
hihifellow posted:Today I had an engineer webex'd in to a computer running an RDP session to a laptop connected to the idrac of a server through a crossover cable so we could get console access. I worked for a while at a company that manufactured a small linux device that had a bunch of cell modems in it that could be used for video broadcasting over IP. The management system was this weird front end that you could log into and send simple bash commands. You could also upload executable scripts. It was a huge pain, since the front end was garbage. So I wrote a small script that I could upload that would open an ssh tunnel back to a server I controlled, and then I'd use reverse ssh proxying to ssh back up that connection into the device. Since the media was encoded via FFMPEG I could also use this to port forward all the video data over the ssh tunnel.
|
|
#
?
May 7, 2015 16:16
|
|
|
SaltLick posted:http://www.eskaustin.com/ "East Side King" sounds like a member of a gang, which is appropriate since that site looks like a wall covered in graffiti. go3 posted:have we covered that this new design of one extremely long web page is loving horrible I don't think "one long page" is a problem at all. I mean, how long have we (internet users in the general sense) been complaining about unnecessary pagination? When an article could fit on one page with some scrolling but instead is split into 5 pages with two tiny paragraphs each, so that the site could get more ad impressions. I think the real problem is the way the images are laid out. Giant fuckoff screen-sized pictures, sometimes multiple ones, every other paragraph, spilling over into margins and breaking up the flow of the article almost as completely as the old pagination poo poo did. I find THAT incredibly annoying. But a single page per article? Gimme dat. Edit: The image spanning is probably a consequence of designing for phones. I'm sure it makes a lot more sense with a small, vertically-oriented screen. But on a widescreen desktop monitor, they suck. Che Delilas fucked around with this message at 16:28 on May 7, 2015 |
|
#
?
May 7, 2015 16:25
|
|
|
Race Realists posted:Anyone here (or have ever been) a Help Desk/Tech Support Manager? I just figure that to be a sort of missing link between Network Admin and Helldesk Literally as of today, I am. And like OAquinas said, not necessarily. My job (every place is unique) is going to largely consist of vendor relations, ordering equipment, being the liaison/contact for VIPs that our department services, and escalation for desktop support issues that my team can't resolve on their own. Add in hiring/firing (I hope not), employee evals, approving time off, and all that, and you see where things start to diverge. My team supports a subset of a much larger organization, so any infrastructure and/or enterprise services are handled elsewhere which limits what I get exposed to / am responsible for. Which is good and bad. Luckily a lot of the things we've set up in the last few months run themselves and our desktop replacement processes are pretty streamlined so I plan on using my downtime to study and practice on VMs before (probably) moving on.
|
|
#
?
May 7, 2015 16:52
|
|
|
Network engineer I Requirements: 5 years experience Kiss me arse.
|
|
#
?
May 7, 2015 17:15
|
|
|
Race Realists posted:Anyone here (or have ever been) a Help Desk/Tech Support Manager? I just figure that to be a sort of missing link between Network Admin and Helldesk I am this as well. I'm tier 3 support, but I only manage the three helpdesk guys, not the four tier 2 guys. It's weird, but it works I guess since we're a small department. I'm also doing server admin and network admin tasks, but that's mostly just shadowing and helping the one server admin guy and one network admin guy we have.
|
|
#
?
May 7, 2015 17:48
|
|
|
Gyshall posted:This is hilarious and I can tell that the guys who run this company are ~~~ MAC GUYS 4 LYFE ~~~~~ just by the design of the website. Most major websites have infinite or near-infinite scrolling. In TYOOL 2015 it's standard-op and with most users now sporting Full-HD 1920x1080 displays why not make use of real estate? Gucci Loafers fucked around with this message at 18:07 on May 7, 2015 |
|
#
?
May 7, 2015 18:05
|
|
|
OAquinas posted:Ehh...not really. The Manager bit means that while you're still on the hook for any sufficiently large technical issue (though that varies with environment/company) your job is primarily herding cats. 
|
|
#
?
May 7, 2015 18:13
|
|
|
Colonial Air Force posted:It comes at the cost of the right salary, but not enough that I mind too much. I want a pic of you skiing in a revwar getup. For...posterity. Yeah....
|
|
#
?
May 7, 2015 18:27
|
|
|
My new job has turned into: Oh poo poo what were these guys thinking- how are we going to support all this legacy crap. To: Oh poo poo, we need to rebuild both HQ's core networks/VM environments, setup two colo's worth of gear, and be ready to build out a HW environment for a new web-based product delivery management system in a few months. Welp. Either time to start drinking or hiring. Guess what's going to happen first.
|
|
#
?
May 7, 2015 18:50
|
|
|
the spyder posted:My new job has turned into: Oh poo poo what were these guys thinking- how are we going to support all this legacy crap. Drink while hiring?
|
|
#
?
May 7, 2015 19:02
|
|
|
luminalflux posted:From what he's saying it kinda makes sense. It just seems like the stupidest thing ever out right, but given that Racklive did the design and build of the holders (so they didn't have to care about that part) and that this has to last just for a depreciation cycle (2-3 years), you can get some good value and money out of it for the time being, especially if Well thats all well in good if you just calculate that out like that, but on EC2 you can do reserved instances and if a machine is shut down you aren't paying for anything but the drive storage. You can also easily switch between memory/cpu/cluster/gpu targeted instances to get the one that is right for what you are doing without having to do much more than shut the machine down and start it back up again (or at the worst create an image and spin up a new one). AWS is always going to appear a little higher cost than dedicated machines but they constantly evolve the system and improve the infrastructure and their offerings and the users don't have to deal with too much hurt from it. Not to mention AWS offers things like hosted databases, load balancers, auto-scaling all of which are easily scriptable. Not to mention the future and form factor of apple hardware is up in the air, so you don't know if you will have to get all brand new mounting racks down the line. I worked for a web host for 5-6 years, and I am just glad I don't have to deal with any of that crap anymore. No more worrying about drive failures, no more overheating machines, no more dealing with customers when you have to RMA their server and do a rush job to put them on something else. Now my current job has turned into making these cool automation stacks that do everything you used to do by hand, and we plan in fault tolerance and security. It pretty much rules. I went from doing hardware grunt work and installing OS, to prepping servers and configuring domains to full of devopsy stacks and I am quite happy.
|
|
#
?
May 7, 2015 22:15
|
|
|
How do you make a switch like that? All the AWS jobs around here want someone with demonstrated experience in ~-devops-~
|
|
#
?
May 7, 2015 22:44
|
|
|
Swink posted:How do you make a switch like that? All the AWS jobs around here want someone with demonstrated experience in ~-devops-~
|
|
#
?
May 7, 2015 23:14
|
|
|
Swink posted:How do you make a switch like that? All the AWS jobs around here want someone with demonstrated experience in ~-devops-~ Mainly just being a good learner and I had a few skills like Mongo and dealing with SOLR already from my last project. I learned to write some python too which might come in handy, but wasn't part of getting the job. Basically, always be learning. If if you aren't learning anything new on a job then its just another job.
|
|
#
?
May 7, 2015 23:55
|
|
|
Do you put your GitHub account on LinkedIn? Your resume? ...should I remove all the VBS from my account? :V Ok so I should push my administration + scripting capabilities. Which are moderate. Cool.
|
|
#
?
May 8, 2015 00:37
|
|
|
Swink posted:Do you put your GitHub account on LinkedIn? Your resume?
|
|
#
?
May 8, 2015 00:58
|
|
|
Last month we started using Slack for chat at my company. This week, my coworker set up a bot in our Ops channel. Ostensibly for useful stuff, but in reality it's just used to interject sick burns and image macros. Current productivity status: zero  Swink posted:Do you put your GitHub account on LinkedIn? Your resume? FWIW I have my github account on my resume.
|
|
#
?
May 8, 2015 01:01
|
|
|
Getting ready to walk out the door and pick up the kids, when we get calls and alarms. For the second time in as many weeks, the file server shares have gotten hit with ransomware  File server itself is fine, but individual users are infected....here's hoping this gives us sufficient ammo to finally remove local admin access
|
|
#
?
May 8, 2015 01:10
|
|
|
Swink posted:Ok so I should push my administration + scripting capabilities. Which are moderate. Cool. Scripting on-prem isn't fundamentally different than scripting in the cloud. Apply!
|
|
#
?
May 8, 2015 01:18
|
|
|
JHVH-1 posted:Well thats all well in good if you just calculate that out like that, but on EC2 you can do reserved instances and if a machine is shut down you aren't paying for anything but the drive storage. You can also easily switch between memory/cpu/cluster/gpu targeted instances to get the one that is right for what you are doing without having to do much more than shut the machine down and start it back up again (or at the worst create an image and spin up a new one). AWS is always going to appear a little higher cost than dedicated machines but they constantly evolve the system and improve the infrastructure and their offerings and the users don't have to deal with too much hurt from it. On this note, has anyone had success with using EC2 as a DR/COOP cold/warm site? One of my customers has zero DR aside from some local backups at their administrative office and at their colo across town. They can't really afford to keep spare hardware stockpiled, but they definitely need the ability to quickly spin up essential services in the event of a major outage or failure.
|
|
#
?
May 8, 2015 01:23
|
|
|
OAquinas posted:Getting ready to walk out the door and pick up the kids, when we get calls and alarms. For the second time in as many weeks, the file server shares have gotten hit with ransomware We got hit with Cryptolocker twice in the past few weeks (and almost got hit a third but it nuked the computer before it could do any damage) and it was kind of annoying to watch Sophos detect the files when we went looking but did nothing to stop it from encrypting everything it could. Today we had our Sophos rep come by, and one of the things he pointed out it is the suspicious behavior detection defaults to "alert, but do nothing" which lets cryptolocker by. Unchecked that option in a hurry.
|
|
#
?
May 8, 2015 01:26
|
|
|
psydude posted:On this note, has anyone had success with using EC2 as a DR/COOP cold/warm site? One of my customers has zero DR aside from some local backups at their administrative office and at their colo across town. They can't really afford to keep spare hardware stockpiled, but they definitely need the ability to quickly spin up essential services in the event of a major outage or failure. People have, but EC2 is just as likely to have problems. If they can't automatically reprovision their stuff if an AZ goes down, it doesn't belong in the cloud, and if they can, then they can already spin stuff up fast. Seriously look at VPSes or "sort-of" cloud solutions like RS/DO which offer a better experience for pets than EC2
|
|
#
?
May 8, 2015 01:30
|
|
|
OAquinas posted:Getting ready to walk out the door and pick up the kids, when we get calls and alarms. For the second time in as many weeks, the file server shares have gotten hit with ransomware That's what I love about my users. They might be terrible with computers, but they don't trust strange messages. Some of them still get viruses, but most of them ask us about anything fishy. Most of the time it's nothing, but I'd rather have five tickets about benign messages than one where we have to root out a virus, let alone one where someone gets most of a site's network infected with cryptolocker. edit: Although I'm pretty sure I get at least one ALL CAPS message per day.
|
|
#
?
May 8, 2015 01:34
|
|
|
OAquinas posted:Getting ready to walk out the door and pick up the kids, when we get calls and alarms. For the second time in as many weeks, the file server shares have gotten hit with ransomware
|
|
#
?
May 8, 2015 01:59
|
|
|
|
| # ? Apr 30, 2024 18:27 |
|
|
Yeah, we were able to contain the damage to a few hundred thousand files and restore from this afternoon, but finding the entrance vector was a bit more troublesome as he had already gone home and his home folder hadn't gotten hit yet. More annoying than anything else.
|
|
#
?
May 8, 2015 02:13
|
|