|
Without amd_pstate=passive:code:code:
|
#
?
Apr 29, 2023 01:58
|
|
|
|
| # ? May 23, 2024 07:03 |
|
|
Mr. Crow posted:Weird, are there errors in the logs? Nothing I could find in the logs, the file organization is roughly... something (they're subfolders divided up within a larger TechTV subfolder but they're not arranged by show but just by group) See though this is the thing, yes I could rename everything to match what jellyfin wants but it's thousands of videos, and then just maybe it would work. They're never going to be well organized because it's not that kind of archive, it's the massive literally everything kind of archive. But this is where it's getting "too much" because, what for? When I really just want to have something aware of folders on an SMB share that can also support playlists with a file browser built in on its own. There doesn't need to be a library, nothing needs to be scanned, they don't need nice names. I just need to be able to know "here are all those videos inside this TechTV folder, I press the top one it starts playing from there" or "here is my (normal easy to edit format like m3u) playlist for TechTV, early 2000s Adult Swim, and bumpers/commercials. I play the playlist it shuffles between it all" It does seem like kawaii player might be it but I'll have to try some things out. a part of the project is off getting repairs so I'll worry about it later.
|
|
#
?
Apr 29, 2023 02:20
|
|
|
VostokProgram posted:Without amd_pstate=passive: Why passive over active? The only time I've touched CPU governors before has been on the Intel side, where the intel_pstate governor only has two settings: "performance" and "powersave", both of which are scaling drivers but let the CPU figure out what to do with less interaction from the OS. Is that what amd_pstate=active is? Twerk from Home fucked around with this message at 02:40 on Apr 29, 2023 |
|
#
?
Apr 29, 2023 02:38
|
|
|
Twerk from Home posted:Why passive over active? The only time I've touched CPU governors before has been on the Intel side, where the intel_pstate governor only has two settings: "performance" and "powersave", both of which are scaling drivers but let the CPU figure out what to do with less interaction from the OS. Is that what amd_pstate=active is? active isn't in 6.2, they only added it in 6.3 e: compare: https://www.kernel.org/doc/html/v6.2/admin-guide/kernel-parameters.html https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html you are right though about what active does: code:Yaoi Gagarin fucked around with this message at 03:02 on Apr 29, 2023 |
|
#
?
Apr 29, 2023 02:52
|
|
|
VostokProgram posted:Without amd_pstate=passive: I get "hardware limits: 2.20 GHz - 4.43 GHz", but ryzen_monitor nevertheless shows the cores idling much lower! OTOH the 4.43 boost is accurate versus a single thread of stress, but IIRC I could get higher than that in windows. Will test out with pstate tomorrow.
|
|
#
?
Apr 29, 2023 03:29
|
|
|
VostokProgram posted:With amd_pstate=passive: Ok, running this now. Ryzen_monitor shows the cores sleeping a whole lot more. Idle power consumption dropped by like 5-8W and idle temp is down 5C. So yes, this does have a real effect and your PC will waste slightly less power. Observed boost is the same though -- was 4.37ghz with one core of stress before, is 4.37 after. I could swear I was getting over 4.4 in Windows the last time I was paying close attention. Which was more than a year ago, so I could be wrong or it could be something on the system has changed. But it also could be that AMD's drivers on Windows identified the cores with the best boost, so 1 & 2 core workloads consistently went there.
|
|
#
?
Apr 29, 2023 18:09
|
|
|
.
Mescal fucked around with this message at 23:40 on Aug 25, 2023 |
|
#
?
Apr 29, 2023 21:01
|
|
|
THE END i got dvd-rs. attempted to burn fedora iso with builtin win10 tool. wouldnt open or begin. got imgburn software, burned iso. chose rewrite sectors on-the-fly so the dvd will load properly or whatever. imgburn finished, popped the drive open, failed to shut the drive, told me to shut the drive and click okay, which i did, then it said what i'll show you in the attached pics. img1io errorcheck conditionstart/stop unit load illegal request invalid field in cdb it was like it could write a disc but not read it. its a pattern. of course it wont work to just use a dvd. attempted to test disc drive with the audio cd. vlc doesnt like it, tells me to refer to the error log, and doesn't add any text to the supposed error log when i tried to print the error log to a txt. and as im testing the dvd drive itself, now it doesnt want to eject. it was able to a minute ago. the fedora dvd i burned on the win10 laptop seemed not to be recognized on the xubuntu tower btw this other bios in the tower, cmos setup util c1985-2008 american megatrends? now thats a good, navigable bios. i booted the tower w the fedora dvd in it, also a thumbdrive which i thought might or might not still have an iso on it. i changed the boot order 1 removable device2 cd dvd3 hdd when it boots it boots to a black screen with a typing cursor flashing, no response to any keyboard mashingneither fedora dvd nor music cd immediately visible in xubuntu gui (thats whats on the tower rn) but since i had the laptop and the tower hooked up at the same time, i tried tethering them. it sets up easy peasy in windows and theres no setup required on the xubuntu end. so i've got an old tower sucking 4gwifi (idk why 5g stopped working) through an ethernet cable. in other words, my xubuntu install is functional and connected to the internet. i'll probably stick with that until my IT friend sells me that reburb in a month or two and get fedora set up on there. i dont want this goofy 1999-looking xubuntu thing, it has nothing i wanted, but it will be enough. in other other words, the mess i have here is working and this saga is over. thank you for your time. i apologize for my emotional blowup. i apologize for the formatting here, its copypasted. hell of an ugly office all this harware makes.
|
|
#
?
Apr 30, 2023 01:05
|
|
|
Missing pic from dvd attempts https://ibb.co/nLPSz47 edit: i still won't turn down help on why nothing but hdd works for anything on the laptop and why booting is effed etc etc. but it's no longer necessary. edit: i should add SPECIAL thanks to the person who suggested tethering. this monster is probably sucking a lot of elecricity rn lol 3 machines and 3 displays Mescal fucked around with this message at 02:33 on Apr 30, 2023 |
|
#
?
Apr 30, 2023 01:07
|
|
|
Mescal posted:the 1tb thumb drive is this model except it says neheui https://www.amazon.com/LOMYGUS-Compatible-Android-2018-2022-MacBook/dp/B0C2NCYR9F i have 1 broken one and 1 brand new one Hard to say, that feels little bit too cheap. I would test the drive by filling it and then verifying that all the files are intact, preferably with MD5 checksums or something.
|
|
#
?
Apr 30, 2023 01:10
|
|
|
You can use f3probe to do a quick test: https://www.linuxbabe.com/command-line/f3-usb-capacity-fake-usb-test-linux
|
|
#
?
Apr 30, 2023 08:03
|
|
|
Mescal posted:Missing pic from dvd attempts  I wouldn't stress your power consumption too much until you start worrying about the walls spontaneously combusting from the power draw and run cables all over to spread that load for safety E: shoulda made sure to get the Chairman Meow pic in frame fully
|
|
#
?
May 1, 2023 17:48
|
|
|
The only offensive thing in that image to me is the amount of noise that must be in that room lol. Otherwise, pretty nice setup! Edit: And do you find heat a problem? I'm in the middle of moving my workstation/gaming PC behind a wall since I'm wasting so much AC this spring cooling the room it's in.
|
|
#
?
May 1, 2023 17:52
|
|
|
I tried to install waydroid for the 6th time in a year on Fedora, and it actually worked this time. I can finally use android (x86) apps in linux. You can mount a local directory into the android container to make passing files easier. You can also set it up so the apps you install on there will be opened up in their own windows like they are native apps. Really happy with it so far, and now I can use moonreader+ which is far superior to the other native linux e-readers.
|
|
#
?
May 1, 2023 19:24
|
|
|
Nitrousoxide posted:I tried to install waydroid for the 6th time in a year on Fedora, and it actually worked this time. I can finally use android (x86) apps in linux. You can mount a local directory into the android container to make passing files easier. Sounds cool. It looks like it will a few hours to download the init image though. Hope it works tomorrow. Where did you find the moonreader apk? And did you get moonreader to sync positions with the regular android version?
|
|
#
?
May 1, 2023 20:09
|
|
|
VictualSquid posted:Sounds cool. It looks like it will a few hours to download the init image though. Hope it works tomorrow. Sometimes you hit a slow mirror. Canceling and trying again is recommended.
|
|
#
?
May 1, 2023 20:20
|
|
|
VictualSquid posted:Sounds cool. It looks like it will a few hours to download the init image though. Hope it works tomorrow. I had it install gapps and then certified google play. From there you can just use the appstore like it's a regular android phone or tablet. Moonreader syncs fine. I setup my webdav server (nextcloud) to sync progress across devices in the app, but you could also use google drive or dropbox. 
Nitrousoxide fucked around with this message at 20:28 on May 1, 2023 |
|
#
?
May 1, 2023 20:26
|
|
|
That waydroid is already annoying me. I have no network and the network troubleshooting stuff they suggest are obviously utterly disconnected from my problem. They are talking about firewall apps that I don't even have installed. Googling sends me to their issue tracker and I find an issue about networking. And there is someone who followed the guide by installing several firewalls and deactivating them. More importantly 90% of posts are about a solution that worked. Turns out waydroid doesn't work if you have docker installed on the same machine. e: well moonreader works now, which is all I wanted. Actually a higher priority that what I was still using docker for on this machine. VictualSquid fucked around with this message at 22:11 on May 1, 2023 |
|
#
?
May 1, 2023 21:51
|
|
|
VictualSquid posted:That waydroid is already annoying me. Do you have Fedora installed on your machine? If so you already have Podman. It is an excellent replacement for docker if that's the problem.
|
|
#
?
May 1, 2023 21:59
|
|
|
Nitrousoxide posted:Do you have Fedora installed on your machine? If so you already have Podman. It is an excellent replacement for docker if that's the problem. No, I am on manjaro. And I have no idea if switching to Podman would fix this. I just moved that last lonely container to my rpi finally.
|
|
#
?
May 1, 2023 22:13
|
|
|
Anyone know if there's a particularly appropriate multicast address range to assign to "lo"? My use case is I have an application that needs to send datagrams to multiple (local) recipients and I'd rather use multicast than write an application-layer dispatch server. In terms of options: AF_UNIX only supports unicast so I have to use AF_INET (either v4 or v6). My read of RFC 6308 is that it doesn't address this particular case. I'd prefer to avoid administratively scoped ranges since there is other multicast usage on the node and I don't want to have to worry about future expansion. It seems the most appropriate might be source-specific multicast (RFC 4607) with a 127.0.0.1 (::1) source address and a 232/8 (or FF3x::/32) group address. Since I know there won't be any other SSM originating from this node (except possibly for local use) in the future, assigning those entire networks to "lo" shouldn't be problematic. Wasn't sure if there was something else I'm missing though.
|
|
#
?
May 2, 2023 01:46
|
|
|
Is using D-bus an option? I believe it supports broadcasting the equivalent of UDP datagrams to multiple recipients. edit: I mean if there's no way to do it the way you want at least BattleMaster fucked around with this message at 02:11 on May 2, 2023 |
|
#
?
May 2, 2023 01:59
|
|
|
BattleMaster posted:Is using D-bus an option? The AF_INET multicast works fine, I'm just trying to figure out if there's a better approach to selecting a group address other than pulling one out of a hat and crossing fingers.
|
|
#
?
May 2, 2023 02:32
|
|
|
Wouldn't you need multiple interfaces with their own 127/8 address and 232/8 address? IPv6 not dedicating a whole range to loopback was one of the best decisions.
|
|
#
?
May 2, 2023 10:13
|
|
|
BlankSystemDaemon posted:Wouldn't you need multiple interfaces with their own 127/8 address and 232/8 address? If it wasn't clear, the reason I want to use multicast is because if you unicast datagrams to multiple processes listening on the same port, the kernel effectively load balances the datagrams between them, so each process only sees part of the traffic. If you use multicast then both (or more) processes see the entire data stream.
|
|
#
?
May 2, 2023 14:18
|
|
|
Me: that waydroid thing is amazing and enough reason to switch to wayland. I was running it the whole day and everything works as good as before. Wayland this morning: https://bugs.kde.org/show_bug.cgi?id=436318 I did find a link to this pretty cool, though unhelpful, page: https://www.gamingonlinux.com/index.php?module=statistics&view=trends#SessionType-top
|
|
#
?
May 3, 2023 11:58
|
|
|
VictualSquid posted:Me: that waydroid thing is amazing and enough reason to switch to wayland. I was running it the whole day and everything works as good as before. I've just had bad experiences with it. X on the other hand, works just the way I expect it to. I'll be glad to see wayland gain more maturity, but whenever I've tried it, it's been wonky at best.
|
|
#
?
May 3, 2023 12:05
|
|
|
Also, I just found out that recursively deleting too many snapshots make zfs stall embarrassingly. As in all other operations on the fs time out and create fs timeout errors. Operations as in writing or trying to list the snapshots to find out if the command is working slowly or actually stalled. It is only 200 snapshots, that is btrfs tier bullshit. And when I got worried at all other operations generating error dumps, I rebooted the system assuming it had hung. Then it took 10 minutes to mount zfs from cache on boot. I spent that time trying to find a way to skip that step, which seems to require booting a live-usb and chrooting to deactivate the mount. Luckly it somehow managed to boot while I was downloading the iso. second worst (non-dataloss) fs experience, right below the multi day mount-time thing in btrfs which at least had a kernel parameter option to avoid it from grub. Rant on today's linux annoyances over. I hope.
|
|
#
?
May 3, 2023 12:33
|
|
|
VictualSquid posted:Also, I just found out that recursively deleting too many snapshots make zfs stall embarrassingly. As in all other operations on the fs time out and create fs timeout errors. Operations as in writing or trying to list the snapshots to find out if the command is working slowly or actually stalled. There's a long-established way of handling this via ZFS Channel Programs - essentially, a Lua script that allows many administrative operations to be run in the same transaction group. It's documented in the zfs-program(8) manual page. The first example in the manual page pretty much shows how to accomplish what you're trying to do, and I'm pretty sure they were invented to address exactly this problem (the first proposal I saw for it involved recursively deleting snapshots). Incidentally, they get their name from mainframes - where a channel program is the generic name for the software that interacts with the storage subsystem. It absolutely sucks that you somehow manged to avoid hearing about them, though. I've seen it with a few thousand snapshots recursively destroyed, it was painful to an extreme degree.
|
|
#
?
May 3, 2023 14:01
|
|
|
BlankSystemDaemon posted:Creating and destroying snapshots involves a fully synchronous transaction group, so yeah if you're doing them in batches it's gonna take a while - especially if the pool is also handling its regular I/O load. I ended up using an xargs script that deletes one snapshot at a time. Which was fast and caused no noticeable slowdown. But, finding out that using the destroy -r facility in that program is slower and causes more load then invoking destroy 200 times is just baffling. If I wrote that about btrfs it would be taken as a sign that it will never take off. e: I am mostly annoyed because I move from btrfs to zfs specifically to avoid this exact sort of trouble. VictualSquid fucked around with this message at 15:44 on May 3, 2023 |
|
#
?
May 3, 2023 14:28
|
|
|
Zfs on Linux lol
|
|
#
?
May 3, 2023 16:02
|
|
|
Windows iTunes organized my Beatles Anthology double albums as "Anthology 2 (1 of 2)", which I don't like in Music on Console, so I'm trying to consolidate these mp3 files into a single directory. I ran the following code on sample data in another directory and it worked perfectly: for file in *.mp3; do mv $file '1-'$file; done When I try this same code in my actual Anthology directory, it gives the correct output when I substitute mv with echo, but when I actually run the code it says "1.mp3 is not a directory". So what the gently caress gives here? Nothing has changed either in the File System or on Music on Console. I'm having to manually rename these files to "1-03 Blah blah blah'.mp3 so that tracks are organized by disk, which is quite annoying.
|
|
#
?
May 4, 2023 02:17
|
|
|
F_Shit_Fitzgerald posted:Windows iTunes organized my Beatles Anthology double albums as "Anthology 2 (1 of 2)", which I don't like in Music on Console, so I'm trying to consolidate these mp3 files into a single directory. I ran the following code on sample data in another directory and it worked perfectly: at a guess you have spaces or other wonky chars in your filenames that aren't present in your test dir. 1.mp3 is the first file it hit, and the rest failed because it exists.
|
|
#
?
May 4, 2023 02:32
|
|
|
F_Shit_Fitzgerald posted:for file in *.mp3; do mv $file '1-'$file; done Gotta quote your variables.
|
|
#
?
May 4, 2023 03:26
|
|
|
F_Shit_Fitzgerald posted:Windows iTunes organized my Beatles Anthology double albums as "Anthology 2 (1 of 2)", which I don't like in Music on Console, so I'm trying to consolidate these mp3 files into a single directory. I ran the following code on sample data in another directory and it worked perfectly: It might be worth it to use a music organizer like beets or picard. Your script issue is probably an issue with spaces in your filename, you need to use more quotation marks. And you will experience more annoying issues if there are apostrophes or other special characters in your filenames.
|
|
#
?
May 4, 2023 09:42
|
|
|
VictualSquid posted:It might be worth it to use a music organizer like beets or picard. I installed Lidarr for this, and I like it very much. Never used beets or picard, maybe those are better.
|
|
#
?
May 4, 2023 13:57
|
|
|
VictualSquid posted:It might be worth it to use a music organizer like beets or picard.
|
|
#
?
May 4, 2023 15:23
|
|
|
cruft posted:I installed Lidarr for this, and I like it very much. Never used beets or picard, maybe those are better. Beets was a thing I installed and looked at, back when I had just switched over from windows and was looking at the options for tagging & lib management. It's a CLI program where you set up a highly configurable / scriptable config and then just run it on files. The good: it's powerful and can do everything. The bad, and why I didn't end up using it: big up-front cost to do all that configuration, assuming you don't want the defaults. Beets is the only music library manager I could ever use -- normally I do 100% manual organization & management. I'm very particular about it. I could write a beets config and scripts to do everything I want. But the effort to set it all up when my library is already well-organized wasn't worth it. I don't add that many new albums per year these days.
|
|
#
?
May 4, 2023 15:33
|
|
|
I just learned from Richard Yao that there's at least one commercial use of ZFS on Windows, which surprised me quite a bit. Possibly because I haven't been paying attention to it, other than keeping vague track of its integration into OpenZFS, but it's still neat.
|
|
#
?
May 4, 2023 15:52
|
|
|
|
| # ? May 23, 2024 07:03 |
|
|
VictualSquid posted:And you will experience more annoying issues if there are apostrophes or other special characters in your filenames.
|
|
#
?
May 4, 2023 16:14
|
|