|
FireSight posted:It was probably the cursor indicating where text would appear.
|
#
?
Apr 21, 2015 03:29
|
|
|
|
| # ? May 16, 2024 21:44 |
|
|
Mo_Steel posted:A lot of tickets came in. One more company realizing that the cloud is not a panacea.
|
|
#
?
Apr 21, 2015 03:39
|
|
|
AlexDeGruven posted:One more company realizing that the cloud is not a panacea. But the cloud!! Shared resources are a bitch. Hosting your critical poo poo on a shared box is asking for trouble. So far my experience with cloud servers is either a ton of small web heads behind a beefy load balancer during peak season or small bullshit servers that aren't really important for anything.
|
|
#
?
Apr 21, 2015 04:03
|
|
|
AlexDeGruven posted:One more company realizing that the butt is not a panacea.
|
|
#
?
Apr 21, 2015 13:20
|
|
|
That is the best browser extension.
|
|
#
?
Apr 21, 2015 13:27
|
|
|
Prescription Combs posted:But the cloud!! Precisely. My team leader mentioned that if we were a startup now, all of our stuff would be in Amazon/Azure. I wanted to choke him, and only partially because I run all of the AIX environment here.
|
|
#
?
Apr 21, 2015 13:33
|
|
|
myron cope posted:Where I work more or less everything is done in coldfusion. Out of curiosity, how is your commute from the present to 2002?
|
|
#
?
Apr 21, 2015 13:43
|
|
|
razorrozar posted:That is the best browser extension. It starts getting confusing when you can't remember if you have it installed or not.
|
|
#
?
Apr 21, 2015 13:48
|
|
|
Prescription Combs posted:But the cloud!! I loved having the "Why aren't we just putting everything in "the cloud" conversation. Why not just get someone else make our SaaS product magically have 5 9s, it can't be that hard a problem .......
|
|
#
?
Apr 21, 2015 14:03
|
|
|
razorrozar posted:That is the best browser extension.  Yes.
|
|
#
?
Apr 21, 2015 16:14
|
|
|
jre posted:I loved having the "Why aren't we just putting everything in "my butt" conversation I may just keep doing this all day
|
|
#
?
Apr 21, 2015 16:18
|
|
|
A ticket came in from a contract agency, for a federal government client. Ticket :go on site to reimage workstation Resolution: Part 1 - image package has not arrived at client location, rescheduled Part 2 - image package has arrived, go on site, package in locked office, only person with key is out, rescheduled Part 3 - image package available, go on site, attempt to image workstation, hd failing, request replacement, rescheduled Part 4 - replacement hd arrived, go on site, replace hd, attempt to image workstation, imaging disk bad, return, request replacement, rescheduled Part 5 - replacement imaging package arrived, go on site, image workstation, ticket closed If this wasn't a ticket for a federal agency in a secure facility, I would have been able to use my own tools and been done in a day. Instead, this whole process took a month. 
|
|
#
?
Apr 21, 2015 17:44
|
|
|
jre posted:I loved having the "Why aren't we just putting everything in "the cloud" conversation. I work at AWS and I wince every time I hear this. Yeah, we allow for a very flexible and nimble infrastructure. But if all you are doing is rebuilding your data center in VPC with EC2 then you are completely missing the point and you will come to wonder in six to nine months where all the supposed savings is and what all the hoopla is all about. You have to design your environment to take advantage of what the cloud offers: servers are herd, not pets, people.
|
|
#
?
Apr 21, 2015 17:54
|
|
|
I wonder why the Kaspersky server stopped working...
|
|
#
?
Apr 21, 2015 18:41
|
|
|
 Gilok posted:I may just keep doing this all day you as well
|
|
#
?
Apr 21, 2015 19:07
|
|
|
Agrikk posted:I work at AWS and I wince every time I hear this. Yeah, we allow for a very flexible and nimble infrastructure. But if all you are doing is rebuilding your data center in VPC with EC2 then you are completely missing the point and you will come to wonder in six to nine months where all the supposed savings is and what all the hoopla is all about. Have any cloud-migration disaster stories? I can't seem to grasp the purpose of emails used with blue text, red text, bold, underlines and highlights. Sure, if something is important bold is appropriate but anything else just makes everything unreadable.
|
|
#
?
Apr 21, 2015 19:07
|
|
|
Tab8715 posted:Have any cloud-migration disaster stories? His emails are just being patriotic.  So this just in, we're choosing the program our entire business runs on based on how it looks. The president of the company has thrown on very decent solutions to all the problems we have because "It doesn't look nice". So instead we're going to pay an additional 30k to have another program customized to include all the features they the other suites already offered.
|
|
#
?
Apr 21, 2015 19:09
|
|
|
Aren't we still waiting feedback regarding your cleanup job?
|
|
#
?
Apr 21, 2015 19:15
|
|
|
jre posted:I loved having the "Why aren't we just putting everything in "the cloud" conversation. Agrikk posted:You have to design your environment to take advantage of what the cloud offers: servers are herd, not pets, people.
|
|
#
?
Apr 21, 2015 19:17
|
|
|
KoRMaK posted:What is 5 9s? 99.999% uptime I'm guessing.
|
|
#
?
Apr 21, 2015 19:19
|
|
|
KoRMaK posted:What is 5 9s? 99.999% uptime, or less than 5 minutes and 15 seconds of downtime in a year. 99.99% meas your service could be down for 53 minutes, 99.9% means it'll go down for 8 hours and 45 minutes and 99% means you're out for over 3 and a half days. And plain one-nine or 90% means your service is so terrible it's out for over a month over an average year. Hardcore financial services typically go for 8 nines or better, where the service provided is only out for fractions of a second every year. Nintendo Kid fucked around with this message at 19:24 on Apr 21, 2015 |
|
#
?
Apr 21, 2015 19:21
|
|
|
Agrikk posted:You have to design your environment to take advantage of what the cloud offers: servers are herd, not pets, people. That's all very well and good for the SF startups I'm surrounded with that are creating scalable web services etc., but what about businesses operating with relatively static workloads? When my clients come to me asking to put their fileserver in the cloud, I do my best to tell them it doesn't do anything for them a colo space wouldn't, but they just hear "the cloud" and go I NEED THAT BECAUSE MAGIC. Case in point, I have a client for which I migrated their Exchange, fileserver, and DCs from a datacenter to AWS. Sure, it's running on reserved instances, but it's costing them essentially the same as the datacenter was (admittedly this was a lovely-rear end cheap datacenter) and they have no resources to expand without paying more, unlike their VM host that had extra space and RAM and was costing them about the same in terms of power and cooling no matter how many VMs they were running. Other than for my own benefit in getting experience with AWS, Powershell, and DSC scripting, there's no point in me spinning up extra Exchange cluster members or DFS targets automatically on AWS to meet demand because those services don't generally have demand expansion at a rate that requires automatic resource expansion. (Though it WOULD be pretty cool to automate adding Exchange members and I'm sure someone has done it). Anyway, I fully agree with herd not pets being the best use for the cloud, but then I'm wondering what the cloud actually DOES offer for people who want pets. To me it seems like forcing a small-medium business peg into a hyperscale hole, which I'm not really a fan of, not to mention the issues with backing up properly and cheaply and simply trusting that your cloud provider won't lose all your data at which point you have no recourse. So is there actually a reason to go to the cloud for those use cases? (Talking about hosting servers, not cloud services like O365). gilok: I've heard of an excellent book called pounded in the cloud by my own cloud.
|
|
#
?
Apr 21, 2015 19:22
|
|
|
KoRMaK posted:What is 5 9s? The herd, not pets principle means that you design for failure in such a way that if a server gets sick, you kill it and redeploy another one just like it. In the classic data center model, you lovingly un-box your server, rack it and cable it, flash all the firmware, install the OS, assign an IP address, give it a name, patch it, and lovingly care for it for its lifetime. The cloud model is such that you build a mechanism so your application, not your servers, are self healing. For example: Instead of a sturdy web server in a rack with 24/7/365 4-hour support, you deploy two or more load balanced servers and if one gets sideways on you, your auto scaling policy kills it when it fails a health check and redeploys its clone in a minute or two. I have EC2 domain controllers in "min 1 max 1" auto scaling groups that take snapshots every night. If the server hiccups, the health check kills it and deploys a clone and AD re-replicates to it from other DCs. This happens automatically and I only know it happened because the next day I have an alert on a dashboard announcing that an instance was redeployed. No human was ever involved in the process. since every machine points to at least two DCs, nothing was interrupted during the five minutes it took to terminate and redeploy. That is what architecting for the cloud means.
|
|
#
?
Apr 21, 2015 19:40
|
|
|
flexlm is a huge piece of poo poo. That is all.
|
|
#
?
Apr 21, 2015 19:40
|
|
|
Potato Alley posted:That's all very well and good for the SF startups I'm surrounded with that are creating scalable web services etc., but what about businesses operating with relatively static workloads? When my clients come to me asking to put their fileserver in the cloud, I do my best to tell them it doesn't do anything for them a colo space wouldn't, but they just hear "the cloud" and go I NEED THAT BECAUSE MAGIC. In this customer's case I would recommend against moving into AWS (or any other cloud provider) for the reasons you describe. If you have four servers running specific and consistent workloads with no expansion planned or anticipated you will never be able to take advantage of AWS' primary benefits. Speed of deployment doesn't matter if you aren't deploying, and agility doesn't mean anything if workloads are static. If the org grows, then maybe there might be a use case for file replication and exchange clustering for decreased latency to the user and more resilience against failure, but that doesn't seem to be the case here. As an AWS employee I'd recommend staying in a classic server room. Unless they don't have the expertise to manage a server room, and at that point using AWS as a managed hosting provider might make sense, because then there is a cost savings realized in letting AWS handle networking, storage, compute, etc for you and that potentially reduces the need for manpower. Agrikk fucked around with this message at 20:03 on Apr 21, 2015 |
|
#
?
Apr 21, 2015 19:59
|
|
|
Migishu posted:Aren't we still waiting feedback regarding your cleanup job? I'm dead. They have internet in hell, btw, but I can only browse GBS and this thread. It was about what I expected. He came in and asked where all "his" stuff was. I said most of it was probably still in the dumpster. He then went literal dumpster diving and pulled out a few trash bags and went through them. He ended up hanging onto a broken canister vacuum, an assortment of power cables, a shovel handle, and a busted computer headset. Some other stuff. I was just stricken by what a sad visage it was: the man I have to work under, the man I have to take orders from, the man that hired me crouched down in an alley digging through trash in order to hang on to... trash. I held my tongue through all of it. He lectured me about throwing away other people's stuff without asking. He pouted for a bit. Now things are back to normal. Apparently at the conference he and the president went to they bought a bunch of stuff. Sensor heads for our virtual set, a central archival NAS(!), a new closed captioning system. The GM told me about all this. When I asked him about it, I got a bunch of vague grunts and "we haven't made any decisions yet." While he was gone, they fired someone in sales. She threatened the sales manager when he escorted her out. I was tasked with disabling her doorcode, something I have asked for training on numerous times, and had no clue about it. I called/emailed/texted him and got no responses. I asked him again yesterday if he'd show me how to do it. He just shook his head and said "She won't be a problem." I reiterated that the GM told me to learn how to do it and he just shook his head again and said, "It's not his decision on what you know how to do around here." So, yeah, he's still purposefully keeping me in the dark on everything regarding my job, despite management's wishes. So either he or I won't be here much longer. I like it here. Looks like it'll be a battle of attrition.
|
|
#
?
Apr 21, 2015 19:59
|
|
|
larchesdanrew posted:"It's not his decision on what you know how to do around here." You have more self control than I do. I can't imagine someone saying some poo poo like this to me and how I would act after.
|
|
#
?
Apr 21, 2015 20:04
|
|
|
Sickening posted:You have more self control than I do. I can't imagine someone saying some poo poo like this to me and how I would act after. I just keep in my head that I have a wife and children dependant on me having what is likely the only available non-minimum wage IT position in this city. I can't afford to get upset. I'm just pretty numb to it at this point.
|
|
#
?
Apr 21, 2015 20:08
|
|
|
Agrikk posted:The cloud model is such that you build a mechanism so your application, not your servers, are self healing. For example: Instead of a sturdy web server in a rack with 24/7/365 4-hour support, you deploy two or more load balanced servers and if one gets sideways on you, your auto scaling policy kills it when it fails a health check and redeploys its clone in a minute or two. Yup, if an EC2 instance misbehaves, take it out back and shoot it in the head and spin up a new one. Netflix developed their Simian Army, and one of the modules randomly kills instances during business hours so if for some reason it doesn't recover they have people available to look into it and determine why. It's a very neat concept.
|
|
#
?
Apr 21, 2015 20:12
|
|
|
larchesdanrew posted:I just keep in my head that I have a wife and children dependant on me having what is likely the only available non-minimum wage IT position in this city. I can't afford to get upset. I'm just pretty numb to it at this point.
|
|
#
?
Apr 21, 2015 20:15
|
|
|
anthonypants posted:It's definitely a thing you need to address with the GM. Or just use the fired woman's door entry code and set the place on fire. Make sure you have cya documentation of your boss refusing you to train you how to deactivate it.
|
|
#
?
Apr 21, 2015 20:20
|
|
|
Agrikk posted:I have EC2 domain controllers in "min 1 max 1" auto scaling groups that take snapshots every night. If the server hiccups, the health check kills it and deploys a clone and AD re-replicates to it from other DCs. This happens automatically and I only know it happened because the next day I have an alert on a dashboard announcing that an instance was redeployed. No human was ever involved in the process. since every machine points to at least two DCs, nothing was interrupted during the five minutes it took to terminate and redeploy. I would like to know more about this
|
|
#
?
Apr 21, 2015 20:35
|
|
|
Spazz posted:Yup, if an EC2 instance misbehaves, take it out back and shoot it in the head and spin up a new one. Netflix developed their Simian Army, and one of the modules randomly kills instances during business hours so if for some reason it doesn't recover they have people available to look into it and determine why. It's a very neat concept. Don't forget Chaos Gorilla: it takes out whole regions. I was there when they turned off the East Coast. No Netflix customer noticed. It was goddamn amazing. Agrikk fucked around with this message at 20:43 on Apr 21, 2015 |
|
#
?
Apr 21, 2015 20:37
|
|
|
Ah yes, but what if the thing watching for failures falls? rly makes u think
|
|
#
?
Apr 21, 2015 20:39
|
|
|
the littlest prince posted:Ah yes, but what if the thing watching for failures falls? Turtles all the way down. Thanks Ants posted:I would like to know more about this YouTube for AWS and "architecting for the cloud". There are a ton of presentations about this topic and what cloud-ready really means. I'd link to some of the better ones but I'm phone posting and  Agrikk fucked around with this message at 21:06 on Apr 21, 2015 |
|
#
?
Apr 21, 2015 21:02
|
|
|
larchesdanrew posted:I'm dead. They have internet in hell, btw, but I can only browse GBS and this thread. Man, I must live in some sort of small corporate paradise because if this happened to me I am nearly obligated to correct it. It's my responsibility to get it right, and it's my obligation to not allow that kind of poo poo go without calling it out. And I don't have to worry about my job over it. It's all in the pursuit of constructive criticism though.
|
|
#
?
Apr 21, 2015 21:12
|
|
|
Ah no I got that bit, it was more specifically about having the systems in place where you can automatically spin up new DCs and kill off the poorly ones.
|
|
#
?
Apr 21, 2015 21:12
|
|
|
the littlest prince posted:Ah yes, but what if the thing watching for failures falls? Who watches the Watchmen?
|
|
#
?
Apr 21, 2015 21:15
|
|
|
A Virus/malware came in; Some people noticed a spam message addressed from our Marketing mailbox this afternoon, and then some of the business partners... and then some of the customers, whoops! I spent last week making sure AVG Cloudcare was up to date and installed on all machines, and there are still people postponing the initial scan. 
|
|
#
?
Apr 21, 2015 21:30
|
|
|
|
| # ? May 16, 2024 21:44 |
|
|
Gerdalti posted:While we're talking about video conferencing, any recommendations for some cheap-midrange video/audio equipment that can be used via USB or something for Lync (SKYPE FOR BUSINESS) and GoToMeeting?
|
|
#
?
Apr 21, 2015 21:37
|
|
