|
turns out that I need to spend a few hours backing up the config before blitzing it, so btrfs will stay for another dayhifi posted:maybe ou 're running out of inodes if this is the case, then I'll hit this error again very soon. The only remedy taken was a reboot. but I suspect that btrfs was not clearing resources, so I couldn't use my free space. So if I don't format, I will likely hit the error again in a few months.
|
#
?
Aug 2, 2017 04:56
|
|
|
|
| # ? Apr 29, 2024 00:29 |
|
|
'df -i' will tell you. it shouldnt really be over 30%
|
|
#
?
Aug 2, 2017 05:49
|
|
|
freebsd 9 or 10 changed the installer algorithm for inodes/gb and people using vms with small amounts of disk were getting hosed because it was giving them linearly less inodes and when you extract the ports tree and compile something it basically explodes a billion source trees in your fs, or in some cases they couldn't even install freebsd because 8gb of storage gave you like 1k files or something really small that the devs didn't forsee happening.
|
|
#
?
Aug 2, 2017 05:53
|
|
|
why would you limit inodes like that?
|
|
#
?
Aug 2, 2017 05:57
|
|
|
because UFS was designed in 1970 and had fixed inode allocation, something something systemd is evil
|
|
#
?
Aug 2, 2017 06:06
|
|
|
hobbesmaster posted:why would you limit inodes like that? its not limiting it's just the default settings for making a new fs, which sort of complicates things when you use those defaults when you're installing the os. back in the day though you'd have a 4 megabyte hard drive or whatever and it really peeved some jorts when 1 meg of that was the inode table. but recently they probably just picked whatever the storage requirement was and made a educated guess as to what would give a comfortable amount of inodes
|
|
#
?
Aug 2, 2017 06:21
|
|
|
btrfs is a system stable enough for production, the issue is that it needs a lot of janitoring that other filesystems don't need, and for someone as rh i guess it makes sense to not support it. lots of pitfalls the customers can fall in that makes thing behave badly
|
|
#
?
Aug 2, 2017 08:03
|
|
|
hifi posted:'df -i' will tell you. it shouldnt really be over 30% /dev/mmcblk0p3 inodes: 0 iused: 0 ifree: 0 iuse% - Mounted on / Same values for the other btrfs partition on /home and the vfat partition on /boot/EFI Maybe btrfs doesn't really have those inode limits that affect other oses?
|
|
#
?
Aug 2, 2017 08:05
|
|
|
oh i guess you 're supposed to use the btrfs tools to get inodes
|
|
#
?
Aug 2, 2017 08:10
|
|
|
rhel is deprecating btrfs this sucks cause my home partition is btrfs
|
|
#
?
Aug 2, 2017 10:50
|
|
|
Condiv posted:rhel is deprecating btrfs so you'll have to come up with some alternative by 2028 god i love properly supported software
|
|
#
?
Aug 2, 2017 11:00
|
|
|
Tankakern posted:btrfs is a system stable enough for production Tankakern posted:the issue is that it needs a lot of janitoring that other filesystems don't need. Lmbo https://www.youtube.com/watch?v=ueZ6tvqhk8U&t=18s
|
|
#
?
Aug 2, 2017 14:36
|
|
|
so apparently my filesystem needed balancing, but it needs free space to do it, so I added a usb stick to the fs to do the balance before removing it now I'm running a cron job to balance the filesystem weekly this reminds me of the time that we had to defragment hard drives as part of proper pc maintenance, except we don't need to do it anymore, because our filesystems are advanced enough
|
|
#
?
Aug 2, 2017 17:13
|
|
|
el dorito posted:so apparently my filesystem needed balancing, but it needs free space to do it, so I added a usb stick to the fs to do the balance before removing it guess u missed the memo where it's a better fs
|
|
#
?
Aug 3, 2017 22:41
|
|
|
ext4eva
|
|
#
?
Aug 4, 2017 01:00
|
|
|
redhat deprecating btrfs does not mean it wont work, or that itll be removed from the kernel, just that red hat isnt going to help you if your poo poo breaks with btrfs filesystems i have been using it on all of my machines for several years and have no plans to stop (btrfs snapshot + send/receive to a backup server owns) but plural of anecdote != data, etc. e: though they could of course patch it out of rhel kernels; itll still be in the upstream kernels
|
|
#
?
Aug 4, 2017 01:02
|
|
|
Yes let's run a custom kernel on an OS that the only reason for installing is being able to CYA with their support
|
|
#
?
Aug 4, 2017 08:14
|
|
|
BobHoward posted:why wouldn't you just download a binary like a normal human being late reply, but gentoo does have binary packages of the most bothersome packages to self-compile, e.g. libreoffice-bin, google-chrome and firefox-bin
|
|
#
?
Aug 4, 2017 10:34
|
|
|
Tankakern posted:btrfs is a system stable enough for production, the issue is that it needs a lot of janitoring that other filesystems don't need, and for someone as rh i guess it makes sense to not support it. lots of pitfalls the customers can fall in that makes thing behave badly redhat is happy to support all kinds of fiddly bullshit. the fiddlier the bullshit, the more "value" they build in your support "subscription" ! they deprecated the btrfs technical preview, because it is not, in fact, stable enough for production. not only that, but they no longer expect it ever will be when poo poo is currently unsupportable but has a future, it enters technical preview. btrfs made it this far when it is finished and meaningfully supportable, it becomes part of the core product. this ain't gonna happen for btrfs, apparently. the future is closed, boys.
|
|
#
?
Aug 4, 2017 15:58
|
|
|
I bet it'll still be in the kernel but they won't support it. I'm sure the kernel has a ton of hardware drivers and iptables rules and esoteric junk that they leave in there with no intention of supporting anyways, but it costs nothing and doesn't increase the attack surface unless the sysadmin decides to load a module I hope they continue to ship the btrfs tools too (competition is nice), but I won't be sad if they don't
|
|
#
?
Aug 4, 2017 21:32
|
|
|
Notorious b.s.d. posted:this is the impossible condition i run tensorflow from the google-compiled binary and every time it runs it says 'hey i wasnt complied withsse4.1/sse4.2/avx/avx2/fma support, think about it' and ok im not doing a ton of work on the cpu but there's probably a little juice to be squeezed out of it (not that im gonna b/c compiling tensorflow is a nuisance)
|
|
#
?
Aug 4, 2017 22:41
|
|
|
pram posted:cool linux stuck on ext until the heat death of the universe honest question how different is ext4 / ext3 / ext2 from each other
|
|
#
?
Aug 4, 2017 22:41
|
|
|
fritz posted:honest question how different is ext4 / ext3 / ext2 from each other ext3 has journaling, which is a big deal ext4 has extents, which is nice
|
|
#
?
Aug 4, 2017 22:50
|
|
|
el dorito posted:ext3 has journaling, which is a big deal ext4 has native xattrs as well. Which is awesome.
|
|
#
?
Aug 5, 2017 02:40
|
|
|
saw this in the f27 change list https://fedoraproject.org/wiki/Changes/NoMoreAlpha cool
|
|
#
?
Aug 5, 2017 06:35
|
|
|
was expecting that to be about the cpu
|
|
#
?
Aug 5, 2017 08:51
|
|
|
quote:may attract more target users (developers) to Fedora this is the biggest issue with fedora rn imo
|
|
#
?
Aug 5, 2017 19:37
|
|
|
i dont know why anyone wants developers as a target user base, they suck
|
|
#
?
Aug 5, 2017 19:38
|
|
|
getting the arch linux crowd on rawhide is probably a cat 6 migraine for fedora devs but maybe it speeds up releases, who knows
|
|
#
?
Aug 5, 2017 19:44
|
|
|
Suspicious Dish posted:i dont know why anyone wants developers as a target user base, they suck 
|
|
#
?
Aug 5, 2017 20:22
|
|
|
apparently, gedit is not being maintained anymore. microsoft kills mspaint, the entire world is in an uproar gnome decides not to support a graphical text editor, probably a lot of shrugging
|
|
#
?
Aug 5, 2017 23:41
|
|
|
Suspicious Dish posted:i dont know why anyone wants developers as a target user base, they suck developers are pretty much the only people who actively want to use linux on the desktop
|
|
#
?
Aug 6, 2017 00:53
|
|
|
el dorito posted:apparently, gedit is not being maintained anymore. the best part of his post is that it is just a thinly veiled advertisement for his incomplete reimplementation of gedit in a library he is bragging about abandoning an application with actual users and complaining about "unnecessary" features while working on a library with neither users nor features peak gnome 3
|
|
#
?
Aug 6, 2017 00:57
|
|
|
el dorito posted:apparently, gedit is not being maintained anymore. how much "maintenance" does an f'ing text editor need? it runs and edits text badabing badaboom its done
|
|
#
?
Aug 6, 2017 01:01
|
|
|
Shinku ABOOKEN posted:how much "maintenance" does an f'ing text editor need? it runs and edits text badabing badaboom its done gnome 3 has no stable apis. not even gtk 3, itself, is stable every release requires app authors to rewrite anything that talks to a gnome api
|
|
#
?
Aug 6, 2017 01:02
|
|
|
I'd consider taking over maintenance of gedit. Mind you, I'd start by rolling back every UI change since gnome2.
|
|
#
?
Aug 6, 2017 01:03
|
|
|
I'm struggling to care about one of the approximately 69,420 half-assed linux text editors becoming unsupported. just pick vim or emacs or pretty much anything that isn't a cadt product and you'll be fine
|
|
#
?
Aug 6, 2017 01:14
|
|
|
Lol if your editor isn't nedit
|
|
#
?
Aug 6, 2017 02:09
|
|
|
RFC2324 posted:Lol if your editor isn't nedit 
|
|
#
?
Aug 6, 2017 02:11
|
|
|
|
| # ? Apr 29, 2024 00:29 |
|
|
only reason i know about nedit was a request from some ee to make it work in kde i couldn't figure out why he wanted it
|
|
#
?
Aug 6, 2017 02:46
|
|