|
i am a moron posted:Anyone else wake up, realize they aren't doing poo poo today (and wouldn't have contributed much if you did?), and be happy about it? I honestly don't know why the companies I work for/with even stay open in December. Same except but woke up by pager duty multiple times for gov cloud and then several more times because of some pretty questionable nightmares and not happy about it
|
#
¿
Dec 16, 2020 16:00
|
|
|
|
| # ¿ May 9, 2024 08:49 |
|
|
i am a moron posted:oh my loving god we're merging with a public company ask for a fat equity grant so you are motivated to share in the company's success
|
|
#
¿
Dec 16, 2020 20:40
|
|
|
Our security org is making several really stupid changes that dramatically increase risk and frustration for zero benefit to appear as though they're increasing security in light of the solarwinds fiasco.code:Live in Europe? Hope you like 11pm releases or 6 am releases Good job guys Methanar fucked around with this message at 22:23 on Dec 17, 2020 |
|
#
¿
Dec 17, 2020 22:19
|
|
|
CLAM DOWN posted:Get Owned You're probably not even joking. this is what security people actually think it means to contribute value.
|
|
#
¿
Dec 17, 2020 22:22
|
|
|
I got paged 4 times and smoke alarm woke up once last night Somebody actually manually created a pd to wake me at 3 am because their dev docker registry wasn't working. What.
|
|
#
¿
Dec 21, 2020 17:37
|
|
|
Biowarfare posted:have you considered just hitting ack and ignoring it because you don't have a 3am response slo for dev The timezone of PD is set to east coast for some reason. It was 2:43 AM.  I did just that. And the psychos immediately escalated as soon as they noticed to somebody else. If they sent it to me again I would have gotten up and asked what the deal is. I was later told quote:dev is dev for our developers, but we really should treat it like production for us. If our dev environment is down at any hour of the day, dozens of engineers are impacted. I ended up getting up at the 4 and 5AM pages anyway because of scary things like `40 pods of important app are unscheduled` and `ingress's conntrack table is full` This whole day has been one giant unending trashfire of problems one after the other. And this is just what I was paged over. 
Methanar fucked around with this message at 00:17 on Dec 22, 2020 |
|
#
¿
Dec 22, 2020 00:15
|
|
|
12 rats tied together posted:It's true actually that dev is just production 2 for ops, but that just means that you should have ops engineers in every timezone that you have developers and have folks oncall outside of business hours. Every ops person we have in Europe is out on PTO right now at the same time lol.
|
|
#
¿
Dec 22, 2020 00:30
|
|
|
Most of the ops issues this morning were because I disabled the cluster autoscaler in a few places on friday after a fire. Something about a combination of unbounded connections being allowed app side on something and a hilarious customer regex that ended up sending us 130x normal traffic on something causing us to saturate an alb, exhaust ephemeral port ranges and cause a kafka rebalance storm. Because I manually did some ASG tweaking as part of the remediation because I was worried about the CA causing a bunch of shuffling to re-optimize binpacking and potentially causing a huge mess of more kafka rebalances at a time we didn't want to. After that friday fire I turned on the CA again but then another ops person told me no we shouldn't lets not rock the boat with causing a bunch of shuffling and re-binpacking on a friday night. I argued that the CA is in place for a reason and leaving it off is more liability than turning it back on. I wasn't in the mood to argue after that friday fire so I relented and left it off. Which was really dumb because I knew it was more liability and if I'm on call: it should be my judgement to do what is less risky since it falls on me when things blow up. This morning the guy who said we should leave the CA off sent me what amounted to `lol we should turn the CA back on`.
|
|
#
¿
Dec 22, 2020 00:46
|
|
|
Biowarfare posted:hey open question to anyone here that is new to kubernetes as an org: this happened to me about 3 months ago because someone was running a singleton
|
|
#
¿
Dec 22, 2020 00:52
|
|
|
12 rats tied together posted:1. Instantly, within the first 20 minutes, at both orgs where we were charged with implementing k8s clusters for the first time. we don't scale kafka brokers. We scale the consumers (95% of everything here is kafka-based one way or another). Its just that our kafka throughput is measured in the petabytes per day with many, many partitions. So shifting around a few hundred consumers all at once can sometimes hurt.
|
|
#
¿
Dec 22, 2020 01:20
|
|
|
uhhhhahhhhohahhh posted:POV: people itt start talking about kubernetes this is now the kafka sensitivities thread post all your kafka stories itt
|
|
#
¿
Dec 22, 2020 01:47
|
|
|
12 rats tied together posted:Lyft's FlinkK8sOperator is worse than just defining a Deployment and rapidly degrades after deployment to the point where the operator state machine just totally breaks and your deployment gets stuck in SubmittingJob or JobCancelled or JobFailed forever, and the only way to recover is to kubectl exec into the job manager, run a cli command to push a savepoint (not a checkpoint) to s3, and then kill it. I don't know anything flinking lyfts but lmao quote:Project Status
|
|
#
¿
Dec 22, 2020 08:24
|
|
|
12 rats tied together posted:I suppose it depends on your business and what your team does day to day -- even having an SLA for a dev environment speaks to a level of disassociation with feature development that is way different than anywhere I have worked in the past (almost entirely SaaS, startups, and enterprises that think they are still startups). It's not that there isn't an SLA necessarily it's just that the SLA is implied: the company writes and runs software. How much do I care about this? A lot -- it is 100% of my job to ensure that the company can continue to successfully write and run software. This is a good post and it's changed my opinion on the matter.
|
|
#
¿
Dec 23, 2020 18:17
|
|
|
i am a moron posted:Frankly I think you’re fostering the us vs. them mentality to fail to empathize with why ops has to ration their time carefully. I get why developers would feel this way about a dev environment, but breaking the walls down would mean compromising on how the usually stretched thin ops teams handle their time the most effectively. That doesn’t usually involve treating dev environments like they are actually generating revenue right this very second and it can’t wait. What does it mean to be a non product-company
|
|
#
¿
Dec 23, 2020 18:19
|
|
|
12 rats tied together posted:"What if we just didn't do anything?"
|
|
#
¿
Dec 28, 2020 00:02
|
|
|
i am a moron posted:By doing this, they’ve acknowledged they see you as a resource instead of a human being. Hiring someone is a crapshoot, and they picked someone with some obscure one-time useful skills that you assuredly could’ve figured out yourself instead of picking someone they know is successful and were (imo) cruelly suggesting should be advancing within your current org. tell the manager to promote you and bring in an SME as a contractor if necessary
|
|
#
¿
Dec 30, 2020 00:24
|
|
|
i am a moron posted:Can anyone explain what possible reason you'd have for not taking the shortest TTLs possible? I don't even understand this convo. I had code review comments recently that told me that 60 second TTLs were too long and I should instead be using 30. I had no reason to disagree. Well-written software should be re-using sockets as often as possible so its not like half the ttl doubles authoritative dns hits. Methanar fucked around with this message at 04:40 on Jan 3, 2021 |
|
#
¿
Jan 3, 2021 04:37
|
|
|
nvm you guys are right I walked into that one.
|
|
#
¿
Jan 3, 2021 05:00
|
|
|
Putting C on your resume is a good way to encourage somebody to interrogate you on it
|
|
#
¿
Jan 5, 2021 05:09
|
|
|
jaegerx posted:I still google loving facls and quotas which is poo poo you don’t need in 2021. Auto mount too. Why is this still on your stupid test? facl you 22 Eargesplitten posted:I just mean like the most basic poo poo for someone that knows how to yum/apt-get/dnf (lol does anything use dnf anymore?)/cat/mkdir and pretty much nothing else. there is nothing else to know
|
|
#
¿
Jan 5, 2021 06:47
|
|
|
22 Eargesplitten posted:I just mean like the most basic poo poo for someone that knows how to yum/apt-get/dnf (lol does anything use dnf anymore?)/cat/mkdir and pretty much nothing else. How do you find the different package versions available for haproxy on a ubuntu system. How do you install a specific version and make sure it doesn't automatically change on you.
|
|
#
¿
Jan 5, 2021 07:32
|
|
|
12 rats tied together posted:Yes, being able to ansible a windows node without having to mass GPO a bunch winrm/psrp nonsense and then figure out a way to locally store AD credentials in a way that doesn't piss off every AD admin to ever exist so I don't need to --ask-pass and type my 36 character full sentence password, which I still have to have and rotate every 90 days for some reason, to manage a win server node would be a good start. 12 rats tied together posted:The only good things Microsoft of the post-2000s has done are WSL and .net core. 12 rats tied together posted:Enter-PSSession is not the same as ssh, it is ~reasonably close in terms of what it allows a human to do, but the important part of supporting ssh is so you can integrate with the entire other half of production server ecosystems which have been using it for the last 25 years. Internet Explorer posted:*whispers* you're not the intended consumer  I have not enjoyed my Azure experiences. It's sort of telling though in response to your several effortposts most of the retort is Clam-level 'nuh uhs' CLAM DOWN posted:Man, you are wrong and ignorant and outdated. Not worth getting into a debate about.
|
|
#
¿
Jan 5, 2021 21:17
|
|
|
jaegerx posted:All cloud providers are poo poo and everyone should run openstack in their own DC. What if I don't want to pretend I work at rackspace though
|
|
#
¿
Jan 6, 2021 02:24
|
|
|
jaegerx posted:I run raid 0 on my gaming pc, AMA. You probably run windows on it too
|
|
#
¿
Jan 6, 2021 02:31
|
|
|
I learned my lesson about trusting the cloud when tumblr decided to start censoring my art
|
|
#
¿
Jan 6, 2021 03:25
|
|
|
My company's bold new diversity initiative for the year is to reallocate the company holidays december 24th and 31st to instead fall on MLK jr and veterans day.
|
|
#
¿
Jan 7, 2021 21:09
|
|
|
I made an actual big boy software architecture recommendation today for some team that is building a new feature in a zoom I got pulled into. That was cool, I don't usually do that. I hope it works lol.
|
|
#
¿
Jan 7, 2021 23:16
|
|
|
DelphiAegis posted:Apparently the full story is even more hilarious. https://www.reddit.com/r/ParlerWatch/comments/kv0jo6/psa_the_heavily_upvoted_description_of_the_parler/ I counter your unsubstantiated anonymous internet post with another unsubstantiated anonymous internet post
|
|
#
¿
Jan 11, 2021 19:43
|
|
|
The Fool posted:I wrote a terraform module on Friday and it’s going in to production today. 
|
|
#
¿
Jan 11, 2021 23:07
|
|
|
jaegerx posted:For you newly remote workers. https://www.amazon.com/dp/B089NFZL9L/ref=cm_sw_r_sms_awdb_imm_t1_qsM.FbZ89HXVB turn your camera off
|
|
#
¿
Jan 13, 2021 04:44
|
|
|
jaegerx posted:Turn your monitor on. that was the joke I was making
|
|
#
¿
Jan 13, 2021 06:54
|
|
|
CLAM DOWN posted:public anger You're so angry. Are you ok (USER WAS PUT ON PROBATION FOR THIS POST)
|
|
#
¿
Jan 16, 2021 21:44
|
|
|
|
| # ¿ May 9, 2024 08:49 |
|
|
If I have a bitlocker encrypted disk and I take an image with dd and write that image back to another disk later. Will I be able to decrypt it normally? I'm pretty sure the answer is yes because I'm not storing anything in a TPM.
|
|
#
¿
Nov 7, 2021 00:14
|
|