|
https://twitter.com/biorhythmist/status/1172378113302355968
|
#
?
Sep 13, 2019 12:30
|
|
|
|
| # ? Apr 26, 2024 17:37 |
|
|
I love the idea that all production metrics and logs should be stored on the same rdbms server as the actual prod data because data should all fit there anyway so that when ops folks start looking a few weeks or months back for incident or perf data, it causes the main site to go slow. either that or I’m conveniently ignoring a bunch of the data stores I have that are not sql because it makes for a good argument.
|
|
#
?
Sep 13, 2019 12:32
|
|
|
but jeff atwood does it!!
|
|
#
?
Sep 13, 2019 12:33
|
|
|
its almost like theres things that rdbms is good at and theres things that nosql is good at and you should know what the right tool to use for the job is? nah, couldn't be.
|
|
#
?
Sep 13, 2019 12:36
|
|
|
our users upload videos but since we only want rdbms and no nosql nonsense, we store all the videos in a blob column. cache churn is killing us, and streaming to customers over the atlantic is really bad. is there bigger hardware should we buy, or is there a dba you know we could hire?
|
|
#
?
Sep 13, 2019 12:40
|
|
|
MononcQc posted:I love the idea that all production metrics and logs should be stored on the same rdbms server as the actual prod data because data should all fit there anyway so that when ops folks start looking a few weeks or months back for incident or perf data, it causes the main site to go slow. i like to write my logs to both sql and elasticsearch. that way when elasticsearch shits itself or gets upgraded with breaking changes or the 3rd party auth solutions stops working you can still read the logs.
|
|
#
?
Sep 13, 2019 12:40
|
|
|
fwiw stack overflow actually runs using replicated sql server clusters so you absolutely wouldn’t need to touch prod to view a log, assuming they log to sql server at all
|
|
#
?
Sep 13, 2019 12:53
|
|
|
yeah I’m assuming they’re not going totally brain dead about it. I’m mostly ranting about the idea that all the data lives in rdbms, which is always false. The file system is a database, cache is a database, S3 is a database, your storage for metrics is a DB, a git server is a database for your code. They’re not RDBMSes and they’re critical to a lot of people’s operations, we just like to pretend that if it’s not in the main DB it doesn’t count. NoSQL is not necessarily scary, t depends on the bounded context you give it.
|
|
#
?
Sep 13, 2019 13:39
|
|
|
Love to scale my loads
|
|
#
?
Sep 13, 2019 14:17
|
|
|
jelql
|
|
#
?
Sep 13, 2019 14:17
|
|
|
interviewing sure is garbage, huh?
|
|
#
?
Sep 13, 2019 14:30
|
|
|
AND recruiting!
|
|
#
?
Sep 13, 2019 14:31
|
|
|
Rex-Goliath posted:interviewing sure is garbage, huh? yep. got my first second-round in five years today and i'm catastrophizing like a mofo can't wait to be asked to rotate a binary tree or to design an iterator from scratch or some dumb bullshit
|
|
#
?
Sep 13, 2019 15:02
|
|
|
CPColin posted:AND recruiting! all to get a job you will want to leave
|
|
#
?
Sep 13, 2019 15:48
|
|
|
Munkeymon posted:all to get a job you will want to leave Thread title
|
|
#
?
Sep 13, 2019 15:53
|
|
|
MononcQc posted:yeah I’m assuming they’re not going totally brain dead about it. I’m mostly ranting about the idea that all the data lives in rdbms, which is always false. a couple pages back a dude argued nosql should be your default position, and then a moron came out of the woodwork to argue that his web app is too big and hairy to be backed by an sql server there's a huge difference between deciding that maybe your log-structured debug data doesn't belong in the sql rdbms, and deciding that 20 gke nodes running a distributed database that requires extensive use of time-lagged secondary indices is better than just ... having an acid database with internal indices
|
|
#
?
Sep 13, 2019 17:36
|
|
|
qhat posted:is there like anything at all that you are correct on the first paragraph is okayish
|
|
#
?
Sep 13, 2019 17:40
|
|
|
got though a screen call and got a tech interview setup next week. cant wait to completely gently caress up something easy like a fibonacci sequence!
|
|
#
?
Sep 13, 2019 19:47
|
|
|
Notorious b.s.d. posted:"schemaless" dude. the preferred nomenclature is “trans” 
|
|
#
?
Sep 13, 2019 20:49
|
|
|
Jim Silly-Balls posted:dude. the preferred nomenclature is “trans” lmao
|
|
#
?
Sep 13, 2019 20:56
|
|
|
Notorious b.s.d. posted:a couple pages back a dude argued nosql should be your default position, and then a moron came out of the woodwork to argue that his web app is too big and hairy to be backed by an sql server The mix of nosql and rdbms we have works well for us. ACID and nosql aren't mutually-exclusive, and for some types of problems eventually-consistent queries with strongly consistent entity key fetches work well enough and can easily scale without human intervention. My point is just that sometimes nosql is the right tool to use, especially when something is in the prototype phase If you're going to call me a moron do it because of my shoes tak fucked around with this message at 22:30 on Sep 13, 2019 |
|
#
?
Sep 13, 2019 22:26
|
|
|
tak posted:You're still tearing down a straw man. It's not 1 web app, it's dozens, developed over the course of the last 15 years, all of them actively being sold (whitelabel SAAS) with many different services running on those nodes (most of which have nothing to do with storage). you were the one who set it up you insisted you had a web app with scale incompatible with the use of an sql rdbms, and when i asked for details, you moved the goalposts a bunch tak posted:My point is just that sometimes nosql is the right tool to use, especially when something is in the prototype phase prototyping is the worst time to use an inflexible tool with complex semantics the "prototype phase" is the point at which you want maximum flexibility and as few unnecessary complications as possible -- i.e. a traditional rdbms tak posted:If you're going to call me a moron do it because of my shoes good shoes won't dig you out of this hole
|
|
#
?
Sep 13, 2019 23:12
|
|
|
Omg shut up about databases u nerds
|
|
#
?
Sep 14, 2019 01:25
|
|
|
Mao Zedong Thot posted:Omg shut up u nerds
|
|
#
?
Sep 14, 2019 01:31
|
|
|
Mao Zedong Thot posted:Omg shut up about databases u nerds not a culture fit
|
|
#
?
Sep 14, 2019 01:47
|
|
|
hobbesmaster posted:not a culture fit 
|
|
#
?
Sep 14, 2019 01:56
|
|
|
Notorious b.s.d. posted:
Complex semantics such as "idempotent APIs" and "stateless services". It's not hard or inflexible. Not really sure what you think nosql is tbh What rdbms do you use?
|
|
#
?
Sep 14, 2019 02:59
|
|
|
tak posted:Complex semantics such as "idempotent APIs" and "stateless services". It's not hard or inflexible. Not really sure what you think nosql is tbh for internal book-keeping i have a mess of postgres and cluster mysql (galera) it may amuse you to know that my job every day is to deal with high-performance nosql clusters, in the multiple billions of queries per day range as a major part of that job i spend a lot of time talking people out of using nosql when they could have a much easier time with a standard postgres deployment it doesn't do me any good to spend time training your devs and poo poo for you to have a horrible loving time because you don't really need this
|
|
#
?
Sep 14, 2019 03:20
|
|
|
Notorious b.s.d. posted:for internal book-keeping i have a mess of postgres and cluster mysql (galera) I had a sensible chuckle Interesting. We're a pretty small company in a small city in the Canadian prairies with some huge global customers, and our cloud costs are surprisingly low (to me anyway). And no IT other than the guy who fixes the sales and support phones, everything is managed (gcp) Maybe we're not using the optimal tech, but I can say that I go home frustrated due to lovely office politics and feeling underpaid and underappreciated, and never because I'm feeling under the weeds with tech debt and bugprone hard to maintain unreliable software or constantly fighting with my tools vOv On topic: how do you deal with toxic manipulative upper managers? This guy is really good at hiding his NPD by being remote and exploiting peoples' insecurity and his relationship with the CEO to keep it quiet, but so many great devs have quit in the last year Do I just stay out of it and apply elsewhere?
|
|
#
?
Sep 14, 2019 04:11
|
|
|
Cold on a Cob posted:yep. got my first second-round in five years today and i'm catastrophizing like a mofo i did ok on the interview. b+ i guess. it helped me narrow down the poo poo i am rusty on so at least now i have an idea of what to brush up on otoh the company does not sound like my kinda party so i'm gonna pass regardless of if they make me an offer. i have several other interviews in the pipeline so hopefully one of them goes well but if not i'll take a break, study for a few weeks, then get back in the mix.
|
|
#
?
Sep 14, 2019 04:15
|
|
|
Notorious b.s.d. posted:as a major part of that job i spend a lot of time talking people out of using nosql when they could have a much easier time with a standard postgres deployment dev lead: *glides confidently into conference room* i hear you're the man to see about getting a nosql db allocated. nbsd: *eyes dev lead's shoes carefully* mmm... no. no, you're going to want to use postgres.
|
|
#
?
Sep 14, 2019 04:18
|
|
|
tak posted:I had a sensible chuckle i work for a gigantic multinational corporation. all the technology choices are a disaster, and i see it as my role to prevent people from choosing self-abnegating choices re: nosql office politics are always lovely, but at least i get paid. (that is pretty much the only thing i care about at work) tak posted:On topic: how do you deal with toxic manipulative upper managers? This guy is really good at hiding his NPD by being remote and exploiting peoples' insecurity and his relationship with the CEO to keep it quiet, but so many great devs have quit in the last year toxic people in upper management reflect the company's overall goals a fish rots from the head
|
|
#
?
Sep 14, 2019 04:19
|
|
|
tak posted:small company in a small city in the Canadian prairies oh my god like I was going to complain about the derail but Jesus Christ I sincerely hope you can escape from camrose or canora or whatever lovecroftian horror of a little town you're trapped in before it eats you alive
|
|
#
?
Sep 14, 2019 06:20
|
|
|
DuckConference posted:oh my god Nah it's not all horrible here. Mortgage is low, my commute is 20 minutes if there's traffic, I can go to my cabin at a later a few hours away whenever I want (pretty much every weekend in the summer), I take 2 vacations to the equator each winter to make the 7-8mo of cold dark bleak gray go by quicker (it really sucks when the only daylight is when I'm at work) I would consider moving to a bigger city but it would have to come with significant figgies and vacation time I put and feelers out a while back and the big remote companies hiring here that I met with pay well but they only give 4weeks vacation max Pretty weird considering I've literally never gotten a bonus nor do I know anyone who's ever got one I think it's a Canadian thing
|
|
#
?
Sep 14, 2019 08:27
|
|
|
Notorious b.s.d. posted:toxic people in upper management reflect the company's overall goals NBSD is quite right on non-shoe matters. i often wonder why our management sucks so much, and then remember that it's because they prioritize providing comforting stats to execs above all else. in unrelated news, the cloudflare IPO has granted me a significant amount of paper wealth! by significant, i mean "approximately equal to the amount of wages i made by taking a job elsewhere". had i staid on longer and completed my vesting period (to get about double the stock i currently have) i would have *lost* money versus what i currently make in wages. startup equity: not even once.
|
|
#
?
Sep 14, 2019 08:59
|
|
|
figuring out what your manager [and their upstream managers] actually want is key because it’s frequently not “deliver the best possible product on time and on budget.” the worst professional relationship I’ve ever had was with a direct manager who didn’t give a poo poo about our team’s output, she just wanted to be the best rated team in the company according to the hr ranking system. I didn’t learn this until after I got fired after a super high scoring review. it helped things make sense in retrospect. the problem was she never told any of us on the team that’s what she wanted, to us it looked like she just kept hiring people then firing them no matter how good their actual work was qirex fucked around with this message at 09:21 on Sep 14, 2019 |
|
#
?
Sep 14, 2019 09:13
|
|
|
tak posted:Pretty weird considering I've literally never gotten a bonus nor do I know anyone who's ever got one nah I’ve received bonuses everywhere i worked (9 years calgary, 9 years toronto area) usually 10% sometimes more
|
|
#
?
Sep 14, 2019 12:02
|
|
|
Cold on a Cob posted:nah I’ve received bonuses everywhere i worked (9 years calgary, 9 years toronto area) Welp
|
|
#
?
Sep 14, 2019 18:57
|
|
|
a deece two figgie bonus they call it a Calgary Thanks
|
|
#
?
Sep 14, 2019 23:01
|
|
|
|
| # ? Apr 26, 2024 17:37 |
|
|
Jim Silly-Balls posted:a deece two figgie bonus the hearts are cold in alberta but so is the cash
|
|
#
?
Sep 14, 2019 23:06
|
|