|
ive used rman before. even in the cli and not the gui one
|
# ? May 18, 2016 00:25 |
|
|
# ? Apr 27, 2024 03:23 |
|
YOSPOS BITCH!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
|
# ? May 18, 2016 00:28 |
|
zen death robot posted:The Bible didn't tell them it was OK to set the systems max shared memory size to be absolute maximum a 64 bit OS would ever allow and go hog wild with SGA sizes and then bitch to me that the storage speeds are loving trash because they're swapping the loving DB cache out to disk constantly like a bunch of loving idiots/ Lol
|
# ? May 18, 2016 00:36 |
|
zen death robot posted:I mentioned they could just allow orcale to automatically handle the memory management and they argued with me that there's a "bug" that wouldn't allow them to use more memory than existed on the system. Lmao
|
# ? May 18, 2016 00:38 |
|
Idk is Oracle one of those DB that expects you to start loving around with resource and memory management? I know some of our systems guys do that for Vertica but still end up blocking users out of the clusters when some team runs a huge data load.
|
# ? May 18, 2016 00:43 |
|
zen death robot posted:HP Nonstop SQL nice
|
# ? May 18, 2016 00:46 |
|
zen death robot posted:I mentioned they could just allow orcale to automatically handle the memory management and they argued with me that there's a "bug" that wouldn't allow them to use more memory than existed on the system. me in a room full of postgres people arguing to vacuum full a table with 99.99999% bloat like 140GB vs 16mb
|
# ? May 18, 2016 00:47 |
|
qhat posted:Idk is Oracle one of those DB that expects you to start loving around with resource and memory management? I know some of our systems guys do that for Vertica but still end up blocking users out of the clusters when some team runs a huge data load. for any serious application, yeah, that's my understanding. you can tune the poo poo out of an oracle db if you really want to
|
# ? May 18, 2016 01:14 |
|
zen death robot posted:oh yeah when you've got a huge DB and you're concerned about every bit of performance you can get you need to Is this the sort of thing that sounds interesting because I don't know anything about it orrrrrr
|
# ? May 18, 2016 01:21 |
|
just use postgres because you aren't a dba at a fortune 100. and even if you are, probably still use postgres. h t h
|
# ? May 18, 2016 01:45 |
|
JewKiller 3000 posted:just use postgres because you aren't a dba at a fortune 100. and even if you are, probably still use postgres. h t h just remember it's a multi-value so uh there is no speed advantage to update in place
|
# ? May 18, 2016 01:46 |
|
berkeley db op
|
# ? May 18, 2016 03:07 |
|
zen death robot posted:oh yeah when you've got a huge DB and you're concerned about every bit of performance you can get you need to also oracle on Windows is awfullllll
|
# ? May 18, 2016 04:15 |
|
zen death robot posted:I didn't post the Nonstop SQL thing as a joke, we actually have some customers that demand that kinda poo poo they don't want their data to ever stop
|
# ? May 18, 2016 05:41 |
|
qhat posted:you know you can just save it all to csv and probably load it into an sqlite db in about 5 seconds right business people can't edit it then though!!! or put hosed up references to random sheets in
|
# ? May 18, 2016 07:58 |
|
Oracle for the extreme loss
|
# ? May 18, 2016 09:14 |
|
Valeyard posted:Oracle for the extreme loss
|
# ? May 18, 2016 16:38 |
|
and the weird alternative join notation apart from that its great to develop on
|
# ? May 18, 2016 16:38 |
|
zen death robot posted:I didn't post the Nonstop SQL thing as a joke, we actually have some customers that demand that kinda poo poo is nonstop sql distributed across datacenters? because my favorite in that category is google's f1 which is a sql db with synchronous multi-datacenter replication they use for their ads backend "oh, the us west coast just fell into the sea? whatever, no biggie, we got 3 datacenters left so it's all good. gotta keep those ad dollars flowing!"
|
# ? May 18, 2016 17:46 |
|
sql server can do that also
|
# ? May 18, 2016 17:53 |
|
YeOldeButchere posted:is nonstop sql distributed across datacenters? because my favorite in that category is google's f1 which is a sql db with synchronous multi-datacenter replication they use for their ads backend quote:"oh, the us west coast just fell into the sea? whatever, no biggie, we got 3 datacenters left so it's all good. gotta keep those ad dollars flowing!" the story i know about nonstop machines is a probably non-existent moment when tandem got a weird support call "hi, there was a car bomb outside and our machine has fallen over" "oh. right. i'm not sure what to do about restarting after a bomb" "no, it fell on it's side, it is still running, we want to know if it's safe to move it" tef fucked around with this message at 19:01 on May 18, 2016 |
# ? May 18, 2016 18:59 |
|
zen death robot posted:sounds about right, the only reason to get one is because you never ever ever ever want an outage even in case of nuclear war the old tech reports are rad http://www.hpl.hp.com/techreports/tandem/
|
# ? May 18, 2016 19:05 |
|
http://www.hpl.hp.com/techreports/tandem/TR-85.7.html ahem
|
# ? May 18, 2016 19:11 |
|
"Why Do Computers Stop and What Can Be Done About It?" Lol
|
# ? May 18, 2016 19:16 |
|
i remember reading this a while ago. the key thing to take away from this is that the most common source of errors is fleshy imperfect meatbags and we need to get rid of them asap the dbas mentioned in this thread seem to corroborate that
|
# ? May 18, 2016 19:18 |
|
YeOldeButchere posted:i remember reading this a while ago. the key thing to take away from this is that the most common source of errors is fleshy imperfect meatbags and we need to get rid of them asap yeah but then you're using an ORM, or worse, stuffing machine-generated JSON dumps into mysql/postgres/mongo without regard for internal schema it will then be irreducably complex and youll find your colleagues have become Young DB Creationists
|
# ? May 18, 2016 19:58 |
|
i can see where they're coming from. i've seen db schemas with so much poo poo in them i wouldn't be surprised if they'd been created 6000 years ago lol at them being intelligently designed though
|
# ? May 18, 2016 23:27 |
|
im the key value table in your terrible database.
|
# ? May 19, 2016 01:40 |
|
you know, the one that houses all the critical business logic in undocumented string fields
|
# ? May 19, 2016 01:42 |
|
schemas and migrations are great when you can do them without downtime backfilling is pretty much what always happens in practice anyhow with live data and you end up resolving the ambiguity in the application, as ever
|
# ? May 19, 2016 02:21 |
|
ask your dba about "the table of numbers"
|
# ? May 19, 2016 02:22 |
|
here's a database riddle: how do you do PRODUCT(column), i.e c1*c2*c3*c4...
|
# ? May 19, 2016 02:23 |
|
very carefully
|
# ? May 19, 2016 02:51 |
|
most sql databases are actually good. at least when you compare with the nosql competition
|
# ? May 19, 2016 03:19 |
|
tef posted:schemas and migrations are great when you can do them without downtime yeah that's why I'm that thing
|
# ? May 19, 2016 03:31 |
|
I really don't want to have to learn SQL but I bet I will have to at some point.
|
# ? May 19, 2016 03:49 |
|
sql is great.
|
# ? May 19, 2016 03:52 |
|
tef posted:here's a database riddle: how do you do PRODUCT(column), i.e c1*c2*c3*c4... code:
|
# ? May 19, 2016 04:14 |
|
nice the old fashioned way is to do exp(sum(log(col)))
|
# ? May 19, 2016 04:17 |
|
|
# ? Apr 27, 2024 03:23 |
|
hahahaha oh man i never would have thought of that, and if i saw it in existing code i wouldn't have understood the purpose either. it's perfect
|
# ? May 19, 2016 04:19 |