|
They only recently fixed Excel so that it didn't scroll by full cells making it very annoying to use when people are using it wrong and putting blocks of text in huge cells
|
#
?
Jul 19, 2022 18:23
|
|
|
|
| # ? Apr 26, 2024 20:59 |
|
|
Had an annoying conversation with my security team today, who’s fighting me on giving developer access to repositories where we maintain services we provide to the rest of the company. Things like the o11y tooling, custom certs, security software that we run on k8s clusters, artifactory repositories, terraform repos for environments devs are not service owners for, the configuration and security settings for our machine images… after about 20 minutes of back and forth, me arguing why read access to system state and write access with controls is good, we finally get to the crux of the issue where he doesn’t trust the controls (open policy agents, SAST/DAST scanning) because some users have excess permissions in gitlab and can override approval rules, which obviously breaks all of our controls. The obvious solution is to fix that so we don’t have ineffective controls, but nooooo, gotta restrict everyone from access at all because we can’t trust our controls  I really like our sec team generally and have good relationships with them, but this constant back and forth on giving service owners access to information about the systems upon which their services depend is agony. For me it’s an observability thing before anything else. I think you should be as liberal as possible with access to configuration details (cluster deployments, IAC, etc) so when the complex interconnections between systems inevitably fail, service owners can independently review the components where you may see failures. Be as restrictive as possible with write access directly on the cloud or software in question (I.e. when not going through IAC). And then be as liberal as you can - with sufficient controls in which you have confidence! - about engineering write access to the codebase through which you deploy your infrastructure and services. imo as long as: - you mandate peer review - have policy checks like SAST/DAST or policy agents - the data or service in question is not internally sensitive - you have confidence in your controls If all the above are true, denying engineer’s read/write access to a given codebase codebase is bad policy. Open access wherever possible always promote innovation, self discovery, and understanding. It ultimately makes for better engineers and better engineering organizations by encouraging cross functional system comprehension and contributions. I limit this specifically to code, obviously this doesn’t apply to IAM permissions, or access to database tables. For context - when I first joined I didn’t even have access to the codebase for the software I was supporting! I can’t possibly be effective in my job supporting engineers if I don’t know what it is I’m supporting. The same goes in reverse. Security through obscurity or inconvenience is not effective security. The Iron Rose fucked around with this message at 18:43 on Jul 19, 2022 |
|
#
?
Jul 19, 2022 18:33
|
|
Cat Army 
|
Internet Explorer posted:Yeah, that poo poo is endemic in organizing spaces. At least some of them will walk you through trying to create a Google Voice number, but even then they don't tell you that it's linked to your Google account and you'll want to release it when you're done. There's really no excuse, so many better ways to do that. Even Spoke would be an improvement. Planned Parenthood's gets $1.3 billion dollars a year. You'd think they'd have this sorted. The Fool posted:even local political campaigns have a phone system that you dial in to that manages all of the outgoing calls There are dozens of Planned Parenthood affiliates in this country so that $1.3bil gets stretched pretty thin. The phone banking is typically handled by the political arm which is legally separate from the clinics, and since it is usually political campaigning their ability to use federal dollars is heavily restricted/prohibited which makes it difficult or impossible to share resources. Other affiliates were better off than mine but I don't think we'd have been able to provide this kind of service even if we wanted to when I was working there.
|
|
#
?
Jul 19, 2022 19:03
|
|
|
Pretty much every volunteer org is a scrappy group of people with no resources. And plenty have to do the same non-profit legal dance. It's just not right to be asking volunteers to be using their own phone number. Especially if it's something potentially charged like Planned Parenthood.
|
|
#
?
Jul 19, 2022 19:15
|
|
|
callhub.io is what I was thinking of, of anyone is ever in a place to make a decision about using this kind of service in the future
|
|
#
?
Jul 19, 2022 19:22
|
|
|
Thanks Ants posted:They only recently fixed Excel so that it didn't scroll by full cells making it very annoying to use when people are using it wrong and putting blocks of text in huge cells
|
|
#
?
Jul 19, 2022 20:45
|
|
|
Pissing me off:Microsoft 365 posted:Your sign-in was blocked ...this is the admin account. If MS wants to verify the identity it has MFA enabled.
|
|
#
?
Jul 21, 2022 20:42
|
|
|
klosterdev posted:Pissing me off: That's a risky login block, which by default blocks the user. That mode is tenant managed with password change to access or full block as options. Someone did weird things with that user.
|
|
#
?
Jul 21, 2022 21:09
|
|
|
Update: Learned a colleague was messing around with conditional and limited access policies for that client right before he went on vacation.
|
|
#
?
Jul 21, 2022 22:43
|
|
|
What happened to not touching poo poo if you’re taking time off right after
|
|
#
?
Jul 21, 2022 23:27
|
|
What happened to that being the best time to do anything  I actually solve this problem by doing nothing, ever
|
|
|
#
?
Jul 22, 2022 01:07
|
|
|
klosterdev posted:Update: Learned a colleague was messing around with conditional and limited access policies for that client right before he went on vacation. If they cant access poo poo, they can't break poo poo while your colleague is on vacation! 
|
|
#
?
Jul 22, 2022 03:34
|
|
|
QuickBooks [any] 2020-2022 R5. Super broken because of the new built-in browser replacing IE. gently caress Intuit, forever and always. https://quickbooks.intuit.com/learn-support/en-us/other-questions/quickbook-desktop-2022-crashing-and-stops-working/00/1091517
|
|
#
?
Jul 22, 2022 17:36
|
|
|
Thanks for that, sent straight to the IT guy at our subsidiary. They’re about six months off from being brought into our workday finance.
|
|
#
?
Jul 22, 2022 18:56
|
|
|
poo poo pissing me off: Seeing windows updates for OneDrive, SkyDrive Pro and Skype for Business when I've theoretically uninstalled them from my desktop. They don't appear in add/remove programs, I've verified that Office has a big red X through SkyDrive, I've yanked the folder out of Program Files and yet updates are still getting downloaded and installed. I don't know why this is so irritating to me but there it is.
|
|
#
?
Jul 25, 2022 06:39
|
|
|
oh man, I totally forgot about skydrive
|
|
#
?
Jul 25, 2022 06:51
|
|
|
and groove!
|
|
#
?
Jul 25, 2022 06:51
|
|
|
Agrikk posted:poo poo pissing me off: OneDrive for business and Skype for Business might get reinstalled if whoever set up the office C2R xml hosed up and forgot to update it. OneDrive is a Windows 10/11 component so it will get reinstalled over time regardless of your attempt to remove it(leave it be if you don’t use it, it will only generate more headaches if it’s not there)
|
|
#
?
Jul 25, 2022 12:36
|
|
|
gently caress, I might be losing my side job now two weeks after losing my second IT job. My boss is having so much work piled on him that he isn't getting enough hours in on his self-employed job to make ends meet so he's basically saying "either you pay me enough for the work I'm doing or I quit" which probably means he's going to quit and I'm certainly not taking his job, doubling my work for a pay cut, so the whole department is probably going to dissolve.
|
|
#
?
Jul 26, 2022 17:49
|
|
|
loving dates piss me off. Dates and date handling in databases. Is it DT_DATE? DT_TIMESTAMP? DT_TIMESTAMPWITHPRECISION? Is it yyyy-mm-ddThh:mm:ssZ? is it MM-DD-YY hh:mm:ss? MM-DD-YYYY? All these bullshit formats. If I see quote:[Flat File Source [2]] Error: Data conversion failed. The data conversion for column "BillingPeriodStartDate" returned status value 2 and status text "The value could not be converted because of a potential loss of data.". I'm going to scream. Just give me a BIGINT milliseconds since Epoch and I'll happily do the conversion myself.
|
|
#
?
Jul 27, 2022 22:15
|
|
|
I have the same problem trying to puzzle out SQL to Pi. I eventually just edited the display I was doing to not have the time stamps because gently caress time
|
|
#
?
Jul 27, 2022 22:29
|
|
|
Naramyth posted:gently caress time yep
|
|
#
?
Jul 28, 2022 03:29
|
|
|
Agrikk posted:Just give me a BIGINT milliseconds since Epoch and I'll happily do the conversion myself. Will you though? With all the leap seconds and time zones and special rules and other bullshit?
|
|
#
?
Jul 28, 2022 04:48
|
|
|
KillHour posted:Will you though? With all the leap seconds and time zones and special rules and other bullshit? POSIX time doesn't do leap seconds and UTC is the only timezone
|
|
#
?
Jul 28, 2022 07:02
|
|
|
Yeah, store in UTC and then apply any time zone conversions when you read it back out
|
|
#
?
Jul 28, 2022 09:36
|
|
|
not always a good idea, if you're trying to store times that are relevant to humans
|
|
#
?
Jul 28, 2022 12:27
|
|
|
Ages ago I wrote a script for the finance team. It watches for data arriving in a particular directory from $CrapDB looks at the XML and based on the content either encrypts and uploads it to $ShitCloudService or moves it to $ManualProcessDir. Simple stuff Last month the project manager put a ticket in to change $ManuallProcessDir to $NewDir where it will get picked up by $FinanceScript written by a FinTech person, handle a load of accounting stuff you'd expect $ShitCloud to do for the money we're paying them, and then pass it to $DropDir for me to encrypt and upload. Again pretty simple stuff. Finance devs handle finance tinkering, IT handle encryption and SFTP. Joe (correctly) doesn't want IT to do banking poo poo, the first script is just a legacy anomaly. Both my upload scripts delete the original after a successful upload because that's what they asked for. Only problem is PM forgot to state that the change wasn't supposed to happen for a month, so I just did it, told them it was done, and closed the ticket. This morning at 9AM while I'm still making my first cup of tea, and thinking about I'm going to implement my current massive scripting project, I get ambished by a huge group Teams call, where it turns out that they had just enabled $FinanceScriot and also switched on a change to $CrapDB's output to support how the script works. Unfortunately when they switched on $FinanceScript it immediately churned through the month's worth of built up data in $NewDir, hosed up the XML and threw it into $DropDir where I encrypt upload and delete it per spec. A months worth of unpaid invoices get rejected by $ShitCloud because the PM didn't request a specific date on the job, nobody checked the contents of $NewDir and nobody noticed invoices going missing for a month! Standard stuff but I struggle with context switching, think that old meme about interrupting a programmer, and it's utterly wrecked my focus on what was supposed to be the first full dev day of my new project after I cleared out my ticket queue yesterday. Also $FinanceScript doesn't use an XML parser and instead assumes one tag per line and parses it with string searching and the Trim function. Lum fucked around with this message at 18:46 on Jul 28, 2022 |
|
#
?
Jul 28, 2022 14:46
|
|
|
Lum posted:Also $FinanceScript doesn't use an XML parser and instead assumes one tag per line and parses it with string searching and the Trim function. 
|
|
#
?
Jul 28, 2022 16:04
|
|
|
We also have a vital XML reading thing that converts it to SQL queries. But if the XML is missing a node, it crashes. If the XML nodes aren't in the correct order, it crashes. If the values in the XML node don't conform exactly to the data constraints of the database, it crashes. If the casing of the XML nodes isn't exactly what the thing wants, it crashes.
|
|
#
?
Jul 28, 2022 17:11
|
|
|
Polio Vax Scene posted:We also have a vital XML reading thing that converts it to SQL queries. I had a to troubleshoot a production impacting problem at a customer where the backups of their mission critical database were failing and causing the application go go offline. The problem was that the script that takes the backup was grepping the out output of the backup command, looking for the string "success" it was using stacked grep and awk commands to count the characters and return like the fifth field in the third block of text separated by a space or some nonsense. The customer had recently switched the storage teir from "standard" to "high priority", the command for the backup actually included the storage class in the response, so the additional space and string threw off the delimiters. The script also looked for "success" and "fail" but had no case for return that didn't match, so when the database was quiesced and a snapshot taken, the command wouldn't catch the error and unpause the database causing a production outage some time later, which quiesced the database would write transactions to memory, which should only have been a few milliseconds, but after like 1/2 hour the server would OOM and crash. Since the time was always some random time after a failed backup it was a little tricky to catch, and since it followed the storage tier change the customer thought there was a hardware problem SAN. It was such a sloppy script, the storage probably had an option to return xml or json or something so it's programmatically scriptable, but no one thinks to look at that.
|
|
#
?
Jul 28, 2022 17:26
|
|
|
poo poo not pissing me off: AWS DNS is down for my company. Oh no, now I have to sit at home and play video games. How awful!
|
|
#
?
Jul 28, 2022 18:10
|
|
|
FlapYoJacks posted:poo poo not pissing me off: AWS DNS is down for my company.
|
|
#
?
Jul 28, 2022 18:52
|
|
|
three days of supporting Onedrive in the enterprise and I'm already done with it. What a temperamental pile of poo poo.
|
|
#
?
Jul 28, 2022 19:09
|
|
|
NO you just SET IT UP WRONG gall - Microsoft
|
|
|
#
?
Jul 28, 2022 20:05
|
|
|
Slow af at the office lately. Think I might have entered round 2 of covid so I took a sick day, end up doing more work from home than I have all week. (wife literally as I write this post: "are you still doing stuff for work?")
|
|
#
?
Jul 28, 2022 21:04
|
|
|
Despite my pre-existing conditions and my workplace posting multiple COVID contacts in my building each day, I am required to come in one day every week, also despite the fact that I'm on a national helpdesk in a team that's distributed in multiple states and I never ever provide support face to face "Opportunities to network and form workplace relationships" is the stated reason for wanting people to return to the office, despite the fact that one day a week I will be in a state of paranoid hypervigilance as I devote my mental energy towards avoiding multiple maskless coughing people rather than my work.
|
|
#
?
Jul 28, 2022 23:52
|
|
|
Just don't come in. And if they ask, "oh I came in Monday"
|
|
#
?
Jul 29, 2022 02:23
|
|
|
I sat through the wildest all-hands meeting today. Spent an hour on the usual corporate crap. Same slogans we've been hearing for a while. "FY 23 is the most important year in company history." "Return to office initiative is going great" "We're better together, as proven by <contrived anecdote about learning more about (thing) in one meeting than in six months of remote work>" "Expected to announce record profits for the fifth quarter in a row." "Focused on hiring talent" That kind of thing. But what was wild was that the employees were having absolutely NONE of it. The comment section of the livestream was absolutely bonkers. People openly and publicly calling out the company's hypocrisy. Record profits, but can't afford cost of living raises. Introduction of stack ranking for performance reviews. Massive attrition across the org, but hiring freezes. My favorite was "I'm the only person left on my team. I'm doing the job of four people. You refused to let me hire. I have a job offer for 3x what I'm making now. What are you going to do when I quit? PS this is my resignation." My second favorite directly called out when leadership directly contradicted itself: "You say you're focused on hiring talent, but you just said you've initiated a hiring freeze. Which is it? Which one is the lie?" It was so wild I threw all the questions into a spreadsheet, did some quick analysis, and threw up some pivot tables. Two thirds of the questions were openly frustrated or angry. Over half openly discussed lack of faith in leadership. Nearly half of the questions were about manpower issues. A third of the questions were about compensation, cost of living, or variations of "I don't get paid enough for this poo poo." A third of the questions were about the return to office initiative and how much the employees hate it. (There was significant overlap where many questions touched more than one subject) The best part is this was just the product/engineering all-hands. The company-wide all-hands is next week and will have triple the attendance. If they don't lock down the comments they may have a full on employee revolt on their hands. PS anyone hiring for a Product Manager, Technical Project Manager, or Senior (non-technical) Project Manager?
|
|
#
?
Jul 29, 2022 03:33
|
|
|
Mustache Ride posted:Just don't come in. And if they ask, "oh I came in Monday" Our boss is super awesome so he was telling us how they've been tracking since rto - the execs are tracking both badge swipes to get into the building as well as work laptop network activity being from in office vs at home. Some people got fired for swiping then going home to work and lying Yes, working from home entirely has been going great, but the executives like how our mission control room looks like NASA when it's full of people (bosses words, not mine) SSJ_naruto_2003 fucked around with this message at 03:49 on Jul 29, 2022 |
|
#
?
Jul 29, 2022 03:46
|
|
|
|
| # ? Apr 26, 2024 20:59 |
|
|
SSJ_naruto_2003 posted:Our boss is super awesome so he was telling us how they've been tracking since rto - the execs are tracking both badge swipes to get into the building as well as work laptop network activity being from in office vs at home. 
|
|
#
?
Jul 29, 2022 03:47
|
|