|
The only reason to run rhel is because the poo poo box $10,000 a seat specialist software tool your company is totally dependent on only targets rhel for support and compatibility. You see, because it's stable, and has "enterprise" right there in the name!
|
#
?
Mar 6, 2018 14:59
|
|
|
|
| # ? Apr 28, 2024 21:02 |
|
|
bruh i would much rather be dealing with a rhel 7 shitbox running some godawful industrial automation crap 20 years from now than a windows 7 or 10 shitbox 20 years from now or a winxp shitbox today
|
|
#
?
Mar 6, 2018 15:05
|
|
|
mrmcd posted:The only reason to run rhel is because the poo poo box $10,000 a seat specialist software tool your company is totally dependent on only targets rhel for support and compatibility. You see, because it's stable, and has "enterprise" right there in the name!
|
|
#
?
Mar 6, 2018 18:11
|
|
|
Sapozhnik posted:bruh i would much rather be dealing with a rhel 7 shitbox running some godawful industrial automation crap 20 years from now than a windows 7 or 10 shitbox 20 years from now or a winxp shitbox today not even rhel support lifecycles are this long solaris might be, though
|
|
#
?
Mar 7, 2018 04:41
|
|
|
Notorious b.s.d. posted:not even rhel support lifecycles are this long I'm sure Oracle would support it that long if you paid them.
|
|
#
?
Mar 7, 2018 04:59
|
|
|
MrMoo posted:I'm sure Oracle would support it that long if you paid them. i spoke from experience sun used to promise minimum support periods of ten years. if you wrote a big enough check, actual support periods extended quite a lot. a former employer was still receiving (some) solaris 7 patches as late as 2010
|
|
#
?
Mar 7, 2018 05:01
|
|
|
It may shock you to learn that a lot of places still use windows xp even though it is now extremely unsupported
|
|
#
?
Mar 7, 2018 05:09
|
|
|
no doubt older versions still in support under ludicrous contracts
|
|
#
?
Mar 7, 2018 05:31
|
|
|
when i worked at oracle they were still patching rhel 4 boxes a client was paying a lot of money to support (~4 years after redhat stopped lol)
|
|
#
?
Mar 7, 2018 05:41
|
|
|
Notorious b.s.d. posted:i spoke from experience I worked for a place still getting Solaris 7 support in 2015. Those engineers did everything they could to blame the hardware every time we opened an incident.
|
|
#
?
Mar 7, 2018 06:56
|
|
|
vulkan 1.1 is out
|
|
#
?
Mar 7, 2018 15:59
|
|
|
should I care?
|
|
#
?
Mar 7, 2018 16:14
|
|
|
one of our production devops guys uses init.d scripts on a system with systemd services
|
|
#
?
Mar 7, 2018 16:20
|
|
|
i just run services with an @reboot in the root crontab
|
|
#
?
Mar 7, 2018 16:24
|
|
|
lol god drat it vulkan 1.0: "trying to parallelize rendering across multiple gpus in the gpu driver required a ridiculous amount of bad heuristics in the driver and was opaque as hell and basically impossible to get right unless you were one of the three gamedev programmers in the whole world who had signed NDAs to access both the nVidia driver source code and the NT kernel source code to debug through everything and see where it was messing up. vulkan makes an explicit design decision to present each gpu to the engine programmer separately and make it their responsibility to distribute work" vulkan 1.1: "lol you can drive multiple gpus simultaneously now"
|
|
#
?
Mar 7, 2018 19:56
|
|
|
vulkan 1.1: "volvo sponsorship, 'marf"
|
|
#
?
Mar 7, 2018 20:01
|
|
|
cinci zoo sniper posted:one of our production devops guys uses init.d scripts on a system with systemd services you mean that thing that people said would break when they changed to systemd works seamlessly because of backwards compatibility? cool..
|
|
#
?
Mar 7, 2018 20:13
|
|
|
Sapozhnik posted:lol god drat it the two are not mutually exclusive. just because there's a vk spec for it doesn't mean it's easy, or it even works in practice without extensions. the bad heuristics are just now in the app.
|
|
#
?
Mar 7, 2018 20:15
|
|
|
Suspicious Dish posted:you mean that thing that people said would break when they changed to systemd works seamlessly because of backwards compatibility? cool.. no i mean i have to hunt for his special snowflake bullshit on our prod
|
|
#
?
Mar 7, 2018 20:22
|
|
|
Notorious b.s.d. posted:not even rhel support lifecycles are this long interestingly, at Vintage Computer Festival Pacific Northwest, I talked to someone who works for Sun on Solaris turns out Oracle didn’t lay everyone off and are still supporting Solaris, but only for people who pay them enough, on the hardware they want to support it on
|
|
#
?
Mar 7, 2018 20:29
|
|
|
Tankakern posted:and kde is very aesthetically pleasing, but people can't be found alive saying anything remotely positive in this subforum i'd agree if i were the dude on the box cover of ESET NOD32
|
|
#
?
Mar 7, 2018 20:40
|
|
|
eschaton posted:interestingly, at Vintage Computer Festival Pacific Northwest, I talked to someone who works for Sun on Solaris morale has been hovering around zero since opensolaris died, long before the layoffs started i can't imagine what kind of stock options package or just sheer masochism would keep your man going back to the office every day
|
|
#
?
Mar 7, 2018 21:24
|
|
|
Notorious b.s.d. posted:i spoke from experience wish there was a  site aggregating all of the latest patches for all of the Solaris versions but, well, Oracle
|
|
#
?
Mar 7, 2018 21:37
|
|
|
cinci zoo sniper posted:vulkan 1.1: "volvo sponsorship, 'marf" vømarf
|
|
#
?
Mar 7, 2018 21:39
|
|
|
eschaton posted:wish there was a it makes me crazy that i can't run old software for sol 8 and sol 9 because i am missing the right patchlevels back when solaris was still A Thing i cared about at work i could just grab them off the internal ftp dump or wherever it was the sysadmins stored them for some reason i was too stupid to archive these for posterity, just assuming solaris would be a normal part of life forever
|
|
#
?
Mar 7, 2018 21:42
|
|
|
Notorious b.s.d. posted:morale has been hovering around zero since opensolaris died, long before the layoffs started he works from home at a very nice place he owns and has piles and piles of equipment from every generation and is seriously into SPARC and Solaris, his work is all done via ssh anyway at VCF-PNW he was showing off his RDI BriteLite, a SPARC “laptop” built out of an IPX logic board and a 640×480 active-matrix LCD in like 1992-3 (so between $16K and $20K new)
|
|
#
?
Mar 7, 2018 21:42
|
|
|
eschaton posted:he works from home at a very nice place he owns and has piles and piles of equipment from every generation and is seriously into SPARC and Solaris, his work is all done via ssh anyway rdi made a lot of weird stuff. they also had alpha and hppa laptops the dominant sparc laptop maker was tadpole, of course. they got bought by general dynamics about five minutes after announcing a dual-cpu laptop. apparently that was a smash hit with the defense community, who knows why
|
|
#
?
Mar 7, 2018 21:53
|
|
|
what was the benefit to solaris in the first place? i never understand the rationale for using it in 2018.
|
|
#
?
Mar 7, 2018 22:23
|
|
|
Lightbulb Out posted:what was the benefit to solaris in the first place? i never understand the rationale for using it in 2018. in 2018, there is no reason to start a new project on solaris
|
|
#
?
Mar 7, 2018 22:31
|
|
|
cinci zoo sniper posted:one of our production devops guys uses init.d scripts on a system with systemd services Don't most systemd system still support the "service <blah> start" syntax and redirect to systemctl? Is that going to go away at some point?
|
|
#
?
Mar 7, 2018 23:13
|
|
|
i don't think that's an actual part of systemd, just a shell script included with debian/ubuntu
|
|
#
?
Mar 7, 2018 23:37
|
|
|
it still exists on centos and fedora too, and I prefer it over systemctl, because typing serv<tab> pos<tab> st<tab> is faster than sys<tab>ctl pos<tab>.<tab> st<tab> systemd commands are incredibly shittily designed for easy autocompletion with minimal keystrokes, and you want to use autocompletion because they're long. people often say unix commands have arcane and unintuitive names, and they're right, but they make autocompletion and history search work like loving magic because they're all unique and short, and if your work has you working in the console even 1% of the time it's super easy to appreciate that.
|
|
#
?
Mar 7, 2018 23:51
|
|
|
Tankakern posted:i don't think that's an actual part of systemd, just a shell script included with debian/ubuntu Ahhh, OK. So I'm guessing that shell script will be deprecated and the sad fucks who didn't adopt proper systemd usage are hosed. 
|
|
#
?
Mar 7, 2018 23:55
|
|
|
said script is just an alias for systemctl. what will happen is one day, systemctl will stop working with poo poo spammed in /etc/init.d/ (it works currently with a wrapper IIRC, since there's still tons of poo poo that hasn't migrated), but will probably be deprecated as soon as it can be.
|
|
#
?
Mar 8, 2018 00:02
|
|
|
fun adventures in filesystems: mac os lets you put a slash in filenames, but HFS+ can't let that happen, so it stores it in the file system as a colon linux will happily copy files off an HFS+ drive and onto an NTFS partition, without complaining about the fact they have an illegal character in the name windows refuses to open any folder containing files or folders with a colon in the filename, and will continue to refuse after you rename the offending file seperately, linux (and by extension the windows subsystem for linux) will gladly create a folder and a Folder with the same names with different case on a windows partition the files in the second folder are completely inaccessible to windows as it will simply open the first one instead linux doesn't care and in fact the developers say that "no, it is windows that is wrong" sorting either of these out is a tremendous pain in the rear end
|
|
#
?
Mar 8, 2018 01:56
|
|
|
atomicthumbs posted:fun adventures in filesystems: just throw all your windows machines in a dumpster, easy Pease
|
|
#
?
Mar 8, 2018 02:03
|
|
|
Cocoa Crispies posted:just throw all your windows machines in a dumpster, easy Pease i'd be happy to if I didn't have to 1. use them for useful work and 2. use them for having fun
|
|
#
?
Mar 8, 2018 02:11
|
|
|
atomicthumbs posted:linux (and by extension the windows subsystem for linux) will gladly create a folder and a Folder with the same names with different case on a windows partition this in particular sounds like a possible attack vector
|
|
#
?
Mar 8, 2018 02:54
|
|
|
atomicthumbs posted:fun adventures in filesystems: expanding on that because the why is mildly funy computer classic macos (pre unixification) used colon as the path separator, so paths looked like my fucken floppy:dumb shite:poop from a butt.txt (except no self respecting classic macos user would use a gauche filename extension) therefore, hfs+, a filesystem designed for classic macos, allows / in filenames but not : at its birth, os x could live on the same hfs+ partition as a classic macos install, and apple had to keep classic alive for many years after the launch of os x. redefining hfs+ filename rules to match unix was not realistic. they settled on behind the scenes trickery to make both worlds happy. if you do this on a mac: code:iirc what actually gets stored on the disk in hfs+ dirents is 'poop/butt', because that's still what's legal in the hfs+ on-disk format. for any unix file api, when trying to create or access files with a : in the name, the kernel automagically does the / : transposition to make it work. for classic (carbon) file apis, some layer or other is doing the equivalent if the underlying fs isn't hfs+ its kinda amazing that it works as well as it does, ive been using os x since 10.0 in 2001 and this was never a pain point as far as i can recall
|
|
#
?
Mar 8, 2018 03:38
|
|
|
|
| # ? Apr 28, 2024 21:02 |
|
|
atomicthumbs posted:fun adventures in filesystems:
|
|
#
?
Mar 8, 2018 03:47
|
|
