|
It's definitely fun organizing the data like that, you get a strong sense of "gonna do this RIGHT!!" which is always refreshing in an IT situation because starting from scratch is a rare opportunity. But when it comes time to write queries for it the error of your ways hits full force.
|
#
?
Feb 16, 2024 16:40
|
|
|
|
| # ? May 27, 2024 10:10 |
|
|
ihafarm posted:Take picture(eta: or screenshot) with iPhone, copy/paste, et voilà. Had a vendor send paper inventory lists of devices, including serial number, mac addresses, etc at very low quality, would not provide in electronic format. Multiple OCR attempts with various scanners and applications failed similarly to your experience. Even with non-optimal shot positioning and wrinkled paper my iPhone produced 100% accurate text. Holy poo poo! Also tehinternet posted:Y’all have any advice or pitfalls to avoid for someone new to it? Always do a SELECT…FROM…WHERE using your criteria before executing a DELETE FROM WHERE Ask me why I do this. Agrikk fucked around with this message at 16:52 on Feb 16, 2024 |
|
#
?
Feb 16, 2024 16:45
|
|
|
Yeah, and turn off autocommit. Anything destructive should be in a transaction, and you should commit it manually. Also, if you are thinking about creating a materialized view, take a step back. Something seriously went wrong for you to be in that situation.
|
|
#
?
Feb 16, 2024 16:56
|
|
|
Antigravitas posted:Also, if you are thinking about creating a materialized view, take a step back. Something seriously went wrong for you to be in that situation. Oh man. Trigger warning that one. I spent six months troubleshooting a poo poo database with a customer, on and on, on and on, never getting anywhere. Then they finally sacked their mat view and suddenly everything fell into place, right as rain. Six months I’ll never get back.
|
|
#
?
Feb 16, 2024 17:22
|
|
|
Antigravitas posted:Yeah, and turn off autocommit. Anything destructive should be in a transaction, and you should commit it manually. Goddamn autocommit. Around 2000, in the era where a 15k disk was the absolute bleeding edge of storage performance, our Oracle team reported unnacceptable DB performance and asked our common management for the funding to double the size of their server, which in those days meant a half-rack sized eight-socket bastard of a machine. Our lead operations DBA looked at the size of their DB and immediately called bullshit on the entire idea. He dug deeper into the code and discovered... autocommit. Yes, if you commit on every single transaction, and your redo logs are on a device that tops out at 200 IOPS, you'll see some significant bottlenecks! He showed them how to modify their code to do explicit commits instead, they defined some sane rules for how often to commit, and all the sudden their existing server performed really well. I'm sure that 24 years later, in the era where thousands of IOPS are cheap, this knowledge is gone, like the office soda fountain whose disposable cups once captured my tears.
|
|
#
?
Feb 16, 2024 17:51
|
|
|
It's easy to make a query pull back the data that you need. It's much harder to do so in an efficient manner. Or architect your database that both meets business needs and performs well. If you know someone who can do those things, treat them like royalty. Because they are. I have a very large customer who has no one who respects databases. Every single launch and migration they do is a complete poo poo show because of it. [Edit: and don't get me started on databases for COTS software that is written like poo poo that then gets a bunch of extra stuff slapped on.] Internet Explorer fucked around with this message at 17:59 on Feb 16, 2024 |
|
#
?
Feb 16, 2024 17:54
|
|
|
Internet Explorer posted:If you know someone who can do those things, treat them like royalty. Because they are. Preach. We have one DBA at my current job who doesn't understand the UNIX platform, doesn't speak English worth a drat, and presumably due to those ESL issues, has only two tones for talking to us: arrogant and demanding, or cringingly servile. I put up with all of it and smile because I've also seen her start work on a database that has been an ongoing performance disaster, do nothing I understand for two days, and then install some sort of performance hint that suddenly makes all the problems go away. Prove you're a wizard, and I'll treat you like a wizard.
|
|
#
?
Feb 16, 2024 18:06
|
|
|
We've hit "keep the lights on" this fine Friday afternoon. Time to hit the gym and head home to sit available on my couch.
|
|
#
?
Feb 16, 2024 20:22
|
|
|
Zorak of Michigan posted:I'm sure that 24 years later, in the era where thousands of IOPS are cheap, this knowledge is gone, like the office soda fountain whose disposable cups once captured my tears. In the era of web devs just making GBS threads out garbage, commits take ages anyway since they are abstracted away behind some REST API. DB engines try their hardest to make a COMMIT; actually persist to disk. It follows that a COMMIT; will make the DB do a lot of work, and the more you can batch into the transaction the better. (Once you realise how conceptually similar file systems and relational databases are, you achieve enlightenment) By the way, query planner magic is sorcery. People who can massage the query planner when it makes bad decisions aren't quite right and should be monitored.
|
|
#
?
Feb 16, 2024 20:56
|
|
|
This is a whole other world for me — is there any reading or YT channels you’d recommend to learn more? I’d prefer to be competent and build things that make sense and run wellAgrikk posted:Holy poo poo! Woof. Feel this in my heart.
|
|
#
?
Feb 16, 2024 21:39
|
|
|
$job-3 I had a team storing attachments in a database. All sorts of xml, bmp, docx, pdf, xls. Basically sharepoint but write it to a table on mysql. Replicating that POS failed as often as it succeeded. If it crashed and hosed poo poo up it’d take 5-7 days to copy it and get all the missing data loaded in again. Managed to bring it back down to 20ish hours but goddamn that poo poo still gives me nightmares.
|
|
#
?
Feb 16, 2024 22:40
|
|
|
tehinternet posted:I just messed around with SQL for a project for the first time and it seems intuitive as hell. I have a natural distrust of that feeling and assume that that will only make it easier to gently caress something up. Every company I've worked for has been too generous with the permissions for the production database. As such, my advice is to set up your tools so that it is ABSOLUTELY CLEAR when you're going to be running queries in production. My preferred method is to use the ability that most tools have to color-code your query windows based on the connection they're using. I suggest the most glaringly bright color you can perceive for anything prod. This is especially important because the connection browser is often disconnected from the query window (meaning you can open a query window for production, then go over to the connection pane and highlight a non-prod connection, and your query window will still be running queries in prod).
|
|
#
?
Feb 18, 2024 04:58
|
|
|
I do the same thing with edge/powershell running as my elevated account, they get red-themed to make sure I pay attention.
|
|
#
?
Feb 18, 2024 14:10
|
|
|
https://old.reddit.com/r/sysadmin/comments/1au697u/can_you_leave_a_tape_drive_over_the_winter_in_the/ This is a cool problem and question to have imo.
|
|
#
?
Feb 19, 2024 15:30
|
|
|
Interesting comment on there, some components (if there's rubber involved) go bad from not enough humidity. My assumption would have been that drier is better for long-term storage.
|
|
#
?
Feb 19, 2024 15:39
|
|
|
Good morning! My boss emailed me late last week to disable TLS 1.0 & 1.1 on some servers. Cool, easy enough! Disabling them broke some SQL poo poo, and I had to re-enable them. Lame!
|
|
#
?
Feb 19, 2024 17:44
|
|
|
chocolateTHUNDER posted:Good morning! My boss emailed me late last week to disable TLS 1.0 & 1.1 on some servers. Cool, easy enough! I'm sure those on-call DBAs were thrilled! Which makes me wonder how much I really want to do DBA work because I drat sure don't want to start taking on-call duties again.
|
|
#
?
Feb 19, 2024 17:47
|
|
|
Hughmoris posted:I'm sure those on-call DBAs were thrilled! I don't think we really have those, but we do have an "application specialist" that normally handles these servers (along with my manager). She's out on PTO today (my company does not have today off) but I saw her checking Teams while this was going on anyway. Felt kinda bad, but also that's on her for not shutting off notifications or whatever on her PTO days off. I've been at this company for almost a year at this point, and there are still some areas that are total black holes to me - it's something I'm still getting used to. This company is a fair bit bigger than the one I came from, so things are a bit more silo'ed. These servers are one of those areas, but I knew what probably broke it since the problem only started immediately after I made those changes 
|
|
#
?
Feb 19, 2024 18:18
|
|
|
I've been at my place for nearly 2 years and there are still black holes in the design of stuff around here. Its disconcerting and yet hard to fix sometimes.
|
|
#
?
Feb 19, 2024 18:29
|
|
|
Does SH/SC have a Discord? Or are people actually working?
|
|
#
?
Feb 19, 2024 18:39
|
|
|
Yeah y'all made me double check my calendar for today
|
|
#
?
Feb 19, 2024 18:41
|
|
|
I am in Japan for work, and the 19th is not a holiday here; however, the 23rd is, so we are hoping to have our datacenter move done by then!
|
|
#
?
Feb 19, 2024 21:28
|
|
|
I don't know of many people outside of government and financial workers in the US that have President's day off.
|
|
#
?
Feb 19, 2024 21:33
|
|
|
It is family day in Canada. Our calendar makes sure to note that Canadian offices are observing the holiday while I’m in office and making no extra money. Last day here’s tomorrow.
|
|
#
?
Feb 19, 2024 21:38
|
|
|
I think its basically a bank holiday around here and nothing else.
|
|
#
?
Feb 19, 2024 22:51
|
|
|
On the topic of database performance a pitfall I've run into a few times is that your database engine will cache tables in memory and as long as your tables are small enough to stay cached you won't notice how inefficient your queries are because it's still lightning fast as long as they're in RAM. So you load up a few gigs of test data and run your tests and everything passes with flying colors, so you sign off and put it in production. Fast forward x months/years and your production data has grown to the point where it no longer fits entirely in RAM cache and your performance drops off a cliff and now all your users are mad about it.
|
|
#
?
Feb 20, 2024 17:34
|
|
|
Collateral Damage posted:On the topic of database performance a pitfall I've run into a few times is that your database engine will cache tables in memory and as long as your tables are small enough to stay cached you won't notice how inefficient your queries are because it's still lightning fast as long as they're in RAM. So you load up a few gigs of test data and run your tests and everything passes with flying colors, so you sign off and put it in production.
|
|
#
?
Feb 20, 2024 18:04
|
|
|
That's a beautiful way to cause an outage. Unsure about what exactly git-receive-pack does, I consulted its manpage. quote:The command allows for the creation and fast-forwarding of sha1 refs (heads/tags) on the remote end (strictly speaking, it is the local end git-receive-pack runs, but to the user who is sitting at the send-pack end, it is updating the remote. Confused?) 
|
|
#
?
Feb 20, 2024 18:10
|
|
|
Me on Thursday: okay you have a new laptop, we should be able to get it set up exactly the same, you'll still need to log into the VPN (like six times) but hopefully this laptop will be faster. Here's your login credentials in case you forget. Boss's boss: okay got it Boss's boss at 9am Tuesday: how do I access my files if I don't have to log into the VPN anymore??? --- Like I know this is super common, but come the gently caress on do you just not listen to other people or have some memory issues or what? I have told you the opposite repeatedly just a few days ago!
|
|
#
?
Feb 20, 2024 18:12
|
|
|
There are people who genuinely struggle to remember things, but in my experience there's a much higher number who simply don't feel they owe you the courtesy of paying any attention to what you are telling them, so the information never got to a point where it could be forgotten
|
|
#
?
Feb 20, 2024 18:16
|
|
|
That is correct. That person is a senior manager of some kind, and they often very much feel they don't need to pay attention to what people say, but I am sorry to tell you that nobody pays attention to anything IT says, ever. This is even true between internal groups within the IT department.
|
|
#
?
Feb 20, 2024 18:19
|
|
|
Thanks Ants posted:There are people who genuinely struggle to remember things, but in my experience there's a much higher number who simply don't feel they owe you the courtesy of paying any attention to what you are telling them, so the information never got to a point where it could be forgotten I work with some directors and VPs who are really dead set on the whole "don't give me extraneous information, give me what I need to help you with your situation and then get to the point" style of communication. And a lot of it isn't because they aren't interested or don't care. It's that their compartmentalization or coping mechanisms don't work, their background threads are going to go nuts chewing on miserable and irrelevant but highly engaging details of other people's problems and keep them from actually being in other moments, listening to things that matter
|
|
#
?
Feb 20, 2024 18:39
|
|
|
I really need a wrap it up button for meetings because holy poo poo there’s a lot of chaff
|
|
#
?
Feb 20, 2024 19:12
|
|
|
We've been acquired, so that's going one of two ways
|
|
#
?
Feb 20, 2024 19:49
|
|
|
Vulture Culture posted:As I enter the later part of my career, I'm finding this is caused by the "ruinous empathy" quadrant of the Radical Candor grid a lot more frequently than I expected, keeping all your background threads spinning on the problems of other people around you until the point that you are no longer able to effectively process any new information at all  I was never a VP or Exec but goddamn if when I'm literally walking the office hallway to get a coffee I am often thinking about crap like the most recent MS Graph Permission abuse. If you want people to pay attention be succent and get to the point. Sending out meeting minutes or actually putting some notes in your meeting before sending it out really, really helps.
|
|
#
?
Feb 20, 2024 19:49
|
|
|
Thanks Ants posted:We've been acquired, so that's going one of two ways Well that explains a lot of the bullshit you seem to have been dealing with. Condolences.
|
|
#
?
Feb 20, 2024 19:52
|
|
|
Thanks Ants posted:We've been acquired, so that's going one of two ways It’s going one way. Acquisitions always end in being stripped for parts as the company needs to inflate its balance books to compensate for the debt in acquiring the company
|
|
#
?
Feb 20, 2024 19:59
|
|
|
Gucci Loafers posted:
"No you still need to use the VPN" strikes me as something even a manager should be able to remember without stenciling it on their forehead.
|
|
#
?
Feb 20, 2024 20:02
|
|
|
BIG FLUFFY DOG posted:It’s going one way. Acquisitions always end in being stripped for parts as the company needs to inflate its balance books to compensate for the debt in acquiring the company Look I was trying to be positive!!!
|
|
#
?
Feb 20, 2024 20:13
|
|
|
|
| # ? May 27, 2024 10:10 |
|
|
it's going in the sense of going, going gone
|
|
#
?
Feb 20, 2024 20:58
|
|