|
|
# ¿ Jun 26, 2017 22:24 |
|
|
# ¿ May 2, 2024 11:28 |
|
pr0zac posted:the terminal security gently caress up is the person leaving their screen unlocked uh do you work in the same office as me?
|
# ¿ Jul 19, 2017 14:55 |
|
Meat Beat Agent posted:yeah it's a bit of a stretch isn't it lmao
|
# ¿ Aug 16, 2017 03:34 |
|
Shaggar posted:Turns out media playback and html rendering are pretty core functions of an OS i too remember active desktop
|
# ¿ Oct 23, 2017 03:27 |
|
Shinku ABOOKEN posted:because the management is using it as a performance indicator which is the dumbest poo poo. have fun telling management that people clicked because they have a billion emails to process and not because they are “bad” employees. it is a performance indicator because lol if your employees blind click links/open attachments in emails that vaguely look like they're related to work instead of thinking about what they're doing. if you regularly get phished during exercises you're clearly not learning from the trainings and are a "bad" employee
|
# ¿ Nov 13, 2017 13:22 |
|
mrmcd posted:What I'm saying is if your attitude is "I will scream at users until they STOP BEING WRONG" instead of trying to engineer guard rails and mitigations to well known problems of human behavior, you're never going to make any real progress. at the end of the day if an attacker sends an email looking like it's from the ceo and joe from accounting sends every employee's w2 back, that's on joe engineering can only go so far to mitigate the human factor
|
# ¿ Nov 13, 2017 13:56 |
|
minato posted:makes sense, Santa is an anagram of @nsa w o w
|
# ¿ Nov 29, 2017 01:04 |
|
Father Jack posted:have you noticed a performance hit? we've seen massive performance hits across our entire infrastructure. the fallout from patching meltdown is going to cost us tens to hundreds of thousands of dollars a month just to tell our customers that we're secure against attacks that will most likely never be able to affect our instances.
|
# ¿ Jan 20, 2018 17:10 |
|
we run a bunch of big cassandra clusters in aws and read and write latency has increased by about 20% which is consistent with what tlp benchmarked: http://thelastpickle.com/blog/2018/01/10/meltdown-impact-on-latency.html we've also noticed that we've had to scale some of our services that are cpu bound by 1.5-2x to keep up with the load
|
# ¿ Jan 21, 2018 17:58 |
|
i say we but really it's our platform engineering and product teams. i get to sit back and tell them to fix security problems
|
# ¿ Jan 21, 2018 18:00 |
|
jre posted:What type of instances and what ami ? we bake our own amis based on ubuntu 14 and run them on i2.4xlarge instances e: for cassandra, that is. services vary in terms of instance type
|
# ¿ Jan 21, 2018 18:03 |
|
yes, exclusively
|
# ¿ Jan 21, 2018 18:05 |
|
jre posted:I'd suggest trying a comparative bench with HVM amazon linux vs your own ami to see if the performance drop is the same that's the next step i think but our platform engineering team will have to update a bunch of our cookbooks because we do a lot of janky ubuntu-specific stuff
|
# ¿ Jan 22, 2018 13:09 |
|
jre posted:I’ve a big write heavy column store workload that we‘ve not patched yet, going to need to bench it to see if its as hosed so follow-up: platform engineers tweaked the write_request_timeout_in_ms setting and bumped it to 5s from the default of 2s. combined with 4 additional nodes and we're actually out-performing pre-patch state the theory is that because writes vary in size, processing the smaller payloads increased cpu overhead and caused larger payloads to time out and drop mutations we also noticed a ton of gc which went back to normal when we did a rolling restart of the cluster to update that setting. the garbage collector might have got in a bad state, contributing to the performance issues
|
# ¿ Jan 22, 2018 22:35 |
|
Loucks posted:Are people itt operating under the belief the Strava has 24/7 tracking data? Because that’s not how it works at all. You’ve got to consciously turn on tracking by starting an activity. “Real-time” premium tracking just means that a given activity is updated live rather than uploaded after it has been completed. goons don't know something but still talk about it with an attempt at authority? stop the presses!
|
# ¿ Jan 30, 2018 14:41 |
|
|
# ¿ May 2, 2024 11:28 |
|
Loucks posted:Was just trying to clarify. Am I wrong? Feel free to enlighten me when you’re done with the snark. you were right, other goons were wrong
|
# ¿ Jan 30, 2018 15:25 |