|
Anyone got a suggestion for scheduled backing up of a full CentOS system (minus /proc, /sys etc.) into a file? Tried bacula: overcomplicated and wouldn't backup /boot Just spent about 3 hours on urbackup, which seemed promising but I just couldn't get it to backup my VM. I had the urbackup server on one VM and it could see the other VM and was outputting to an NFS share but the bastard thing would not backup. There must be something that'll run as a client on the guest which uses a master Linux box to orchestrate. I want to backup a running system/systems. Edit: I didn't mean miss out /etc, obviously. I'm just tired and bad syntax. apropos man fucked around with this message at 23:45 on Oct 11, 2018 |
# ? Oct 11, 2018 23:41 |
|
|
# ? Apr 20, 2024 03:40 |
|
kujeger posted:I mean, sure, I hate Exactly. It's not like you hate Mac and its users, it's more that they're irrelevant. All 5 of them. They and their special needs OS cry loudly all day to give you the impression that they're dozens. Meh, let 'em cry.
|
# ? Oct 12, 2018 01:57 |
|
Volguus posted:Exactly. It's not like you hate Mac and its users, it's more that they're irrelevant. All 5 of them. They and their special needs OS cry loudly all day to give you the impression that they're dozens. Meh, let 'em cry.
|
# ? Oct 12, 2018 02:03 |
|
apropos man posted:Anyone got a suggestion for scheduled backing up of a full CentOS system (minus /proc, /sys etc.) into a file? For free?
|
# ? Oct 12, 2018 02:14 |
|
apropos man posted:Anyone got a suggestion for scheduled backing up of a full CentOS system (minus /proc, /sys etc.) into a file? Does it have to be a single file? If not, you could use bup. If the master can ssh into the VM, install it on both and then `bup on <VM> index <paths> && bup on <VM> save /' or similar from the master.
|
# ? Oct 12, 2018 03:05 |
|
Vulture Culture posted:#!/usr/bin/env bash unless you hate Mac users But I do... The right answer is clearly ksh anyway. For backups, I'm a grognard who still swears by amanda evol262 fucked around with this message at 03:51 on Oct 12, 2018 |
# ? Oct 12, 2018 03:43 |
|
jaegerx posted:For free? Not necessarily. I'm trialling stuff at home that's possibly going to be used at work. We went with Acronis Backup briefly, but their RAID driver support for recovery on a Dell PowerEdge sucked. I know that sounds unlikely but, believe me, trying to roll your own drivers into a WinPE recovery disk and then having it not recognise the LTO7 we were storing backups on was frustrating. Our VM's are Linux but the Linux version of the recovery ISO wouldn't recognise the RAID. Then the WinPE version recognised the RAID but not the tape drive. Catch 22. ToxicFrog posted:Does it have to be a single file? If not, you could use bup. If the master can ssh into the VM, install it on both and then `bup on <VM> index <paths> && bup on <VM> save /' or similar from the master. I'll have a look at that. Thanks.
|
# ? Oct 12, 2018 06:31 |
|
apropos man posted:Anyone got a suggestion for scheduled backing up of a full CentOS system (minus /proc, /sys etc.) into a file? The Ceph group at work chose Bareos, a Bacula fork. I think one of their reasons was that it was simpler.
|
# ? Oct 12, 2018 21:45 |
|
Are there any ffmpeg wizards in here, or could you point me to the correct thread? I'm running into an issue with trimmed video files being the correct length but showing incorrect duration in media players. I can go into more details if anyone knows ffmpeg and wants to skim some logs.
|
# ? Oct 12, 2018 23:48 |
|
CRAYON posted:Are there any ffmpeg wizards in here, or could you point me to the correct thread? That's a funny one: I have used the ffmpeg C API to encode/decode/merge/resample poo poo but I have zero knowledge and experience with the ffmpeg console program that pretty much does the same thing. However, from my experience, when the standard video players show the wrong time it just meant that I messed up the frames drainage. That is, the video has a certain metadata and because for whatever reason i didn't write all the required frames ("drain" the encoder) to the output the video players just wing it and show whatever strikes their fancy. But that's not a problem applicable to the ffmpeg console program, which does everything in the correct way.
|
# ? Oct 13, 2018 00:34 |
|
apropos man posted:Not necessarily. I'm trialling stuff at home that's possibly going to be used at work. We went with Acronis Backup briefly, but their RAID driver support for recovery on a Dell PowerEdge sucked. I know that sounds unlikely but, believe me, trying to roll your own drivers into a WinPE recovery disk and then having it not recognise the LTO7 we were storing backups on was frustrating. Our VM's are Linux but the Linux version of the recovery ISO wouldn't recognise the RAID. Then the WinPE version recognised the RAID but not the tape drive. Catch 22. Oh you want a DR solution. That’s not backup. You want to build bootable isos of your install?
|
# ? Oct 13, 2018 02:12 |
|
jaegerx posted:Oh you want a DR solution. That’s not backup. You want to build bootable isos of your install? What we've been using for years is Lonetar (I know, it's ancient), which makes a daily .tar of the whole OS (you can exclude certain dirs) and you make an ISO rescue CD for each system. The ISO has a kernel and partition table on it which is enough to recover the system that was tarred on any particular day. I suppose I'm looking for something similar, and Acronis is supposed to do something similar except it uses a generic recovery ISO. An ISO which is incompatible with our hardware, whether you use the WinPE ISO with Dell PERC drivers rolled into it or whether you use the Linux ISO, which doesn't have the PERC drivers but can read our Ouantum tape drives. Saukkis posted:The Ceph group at work chose Bareos, a Bacula fork. I think one of their reasons was that it was simpler. Another one to look into, cheers.
|
# ? Oct 13, 2018 10:07 |
|
CRAYON posted:Are there any ffmpeg wizards in here, or could you point me to the correct thread? Sure. What command are you using, or is it a GUI? One thing I can think of is the beginning of your segments might not be at keyframes. That would require you to re-encode the entire video, or else weird issues like what you’re describing can crop up. Starting cuts at keyframes avoids that issue and doesn’t result in quality loss.
|
# ? Oct 13, 2018 15:23 |
|
Why is it that none of the operating systems can't deal with multiple disks correctly anymore? When I installed Windows back then, with two SSDs in the system, it tried to install the bootloader to the wrong SSD. Now I tried to install Ubuntu to my system, it attempted the same and luckily failed. It let me chose a different drive after that error, but it didn't do squat with that choice. I actually had to spin up a VM, pass through the SSD and install things in there.
|
# ? Oct 14, 2018 19:30 |
|
UEFI means the mobo decides which disk gets priority.
|
# ? Oct 14, 2018 19:42 |
|
Volguus posted:That's a funny one: I have used the ffmpeg C API to encode/decode/merge/resample poo poo but I have zero knowledge and experience with the ffmpeg console program that pretty much does the same thing. However, from my experience, when the standard video players show the wrong time it just meant that I messed up the frames drainage. That is, the video has a certain metadata and because for whatever reason i didn't write all the required frames ("drain" the encoder) to the output the video players just wing it and show whatever strikes their fancy. Haha, you're on to something. I'm using the ffmpeg-python library. More details below. Double Punctuation posted:Sure. What command are you using, or is it a GUI? Okay so I am doing what I thought was the exact same thing with two different tools, ffmpeg-python and the regular command line tool. The ffmpeg-python tool is producing files that report incorrect duration while the command line tool files are reporting the correct duration. I'm specifically changing the filetype with both tools because I want ffmpeg to re-encode the video so that I can clip down to the millisecond and not have to start at a keyframe. ffmpeg cli (log): code:
code:
|
# ? Oct 16, 2018 16:03 |
|
Is there a command that creates a ring buffer of the last n seconds? I've got a tricky hangup that I can reproduce only so often and I'd like to record the last 60 seconds of an strace without growing the log infinitely large.
|
# ? Oct 20, 2018 20:01 |
|
It depends on how the process that logs writes its output. If it's to a file that the logging process constantly keeps open, you're probably out of luck because you can't remove the head of the file while it's open. If it's to a file that the logging process opens/appends/closes for each line, then use logrotate. And if it's logging to stdout or something similar, you can probably pipe it to a bash script that uses a "while read" loop to log the lines to disk, and then calls logrotate every X lines. E: this could probably be code-golf'd better, but you get the gist. This will pipe the output of a "strace ls -l" into a bash script, and there's a line where you have the chance to run logrotate (but I didn't implement it). Also, look into the "split" command as this can automatically distribute a single input stream over many files, but it might not be appropriate for a ring buffer. code:
minato fucked around with this message at 21:48 on Oct 20, 2018 |
# ? Oct 20, 2018 21:28 |
|
If you have Apache installed, look at the included rotatelogs command. If not, I guarantee you can find a zillion stdout rotating solutions with some google.
|
# ? Oct 20, 2018 21:52 |
|
xzzy posted:If you have Apache installed, look at the included rotatelogs command. Perfect, thanks! Ended up with this which works great: code:
|
# ? Oct 20, 2018 22:48 |
|
I want to make a Vim mapping to put stuff in the mac clipboard. I do this with w !pbcopy in normal mode. How can I make one that works with visual line selection too?
|
# ? Oct 22, 2018 15:01 |
|
I have an Ubuntu 18.04 install that is running out of space on /boot. Currently the drive partitioning scheme is:code:
Resize the lvm partion down and increase the boot partion or I could try to move boot to the root lvm partition and do away with the separate mbr boot partition all together. Which of these is the best idea / least likely to eat my data and how would I do that?
|
# ? Oct 22, 2018 17:11 |
|
You might have old kernel versions hanging out in /boot that you can just delete to reclaim some space.
|
# ? Oct 22, 2018 17:16 |
|
Docjowles posted:You might have old kernel versions hanging out in /boot that you can just delete to reclaim some space. Ah, yeah I should I have been more specific. I have already done that. Basically I am at the point where I have to manually remove the one old kernel before there is space for a new kernel. Also I tried to run the upgrade to 18.10 the other day and it complained of lack of space on boot. Just tired of the constant janitoring of /boot and trying to get it to where I can just sudo apt upgrade and have it work without complaining constantly.
|
# ? Oct 22, 2018 17:43 |
|
The XKCD Larper posted:I want to make a Vim mapping to put stuff in the mac clipboard. I do this with w !pbcopy in normal mode. How can I make one that works with visual line selection too? The + register is also your system clipboard. I’d give you an example, but I honestly can’t think of the actual button presses. It’s all muscle memory. Basically instead of yanking to an alphanumeric register, use +. You can also paste from it.
|
# ? Oct 22, 2018 19:17 |
|
Horse Clocks posted:The + register is also your system clipboard. It's "+yy (yank-to-system) or "+p (paste from system)
|
# ? Oct 22, 2018 19:30 |
|
jaxercracks posted:Ah, yeah I should I have been more specific. I have already done that. Basically I am at the point where I have to manually remove the one old kernel before there is space for a new kernel. Also I tried to run the upgrade to 18.10 the other day and it complained of lack of space on boot. Just tired of the constant janitoring of /boot and trying to get it to where I can just sudo apt upgrade and have it work without complaining constantly.
|
# ? Oct 22, 2018 19:58 |
The XKCD Larper posted:I want to make a Vim mapping to put stuff in the mac clipboard. I do this with w !pbcopy in normal mode. How can I make one that works with visual line selection too? The previously-posted solutions in this thread work great and you should definitely use them, but I've been reading up on vimscript a bit and I thought I'd give this a shot. code:
|
|
# ? Oct 22, 2018 21:19 |
|
Having a separate boot partition for just home installation is a huge hassle: 1. Size will probably be wrong unless much bigger than standards in installers. 2. Windows 10 is just poo poo in general but it creates the EFI without asking about what size you want it, and if you as much as touch it will completely poo poo the bed. 3. Special snowflakes like Solus which make it even more of a hassle to manage several distros.
lordfrikk fucked around with this message at 10:06 on Oct 23, 2018 |
# ? Oct 23, 2018 10:04 |
|
anthonypants posted:How did you remove the old kernels or verify that they were deleted, and what's using space on /boot? Removed with apt the standard apt remove linux-image-blah-blah-blah then periodically clean up orphaned stuff by hand. I don't remember at this point how I ended up with this size boot partition. But it really seems just maybe half as big as needed. Here is my current /boot: code:
Obviously I have recently cleaned this out. So you can see that the current kernel stuff takes up right at 70mb. Which gives you just barely enough space to have three kernels in there. The kernel autoremove script indicates that depending on how it is called you might end up with up to four kernels. So it seems like enough space for four kernels plus a bit of slack would be the correct amount.
|
# ? Oct 23, 2018 15:54 |
lordfrikk posted:Having a separate boot partition for just home installation is a huge hassle: 1. Size will probably be wrong unless much bigger than standards in installers. 2. Windows 10 is just poo poo in general but it creates the EFI without asking about what size you want it, and if you as much as touch it will completely poo poo the bed. 3. Special snowflakes like Solus which make it even more of a hassle to manage several distros. What makes Solus a special snowflake distro? I don't really know much about it other than the fact that it's rolling release and people tend to like it.
|
|
# ? Oct 23, 2018 16:55 |
|
jaxercracks posted:Removed with apt the standard apt remove linux-image-blah-blah-blah then periodically clean up orphaned stuff by hand. I don't remember at this point how I ended up with this size boot partition. But it really seems just maybe half as big as needed. Here is my current /boot:
|
# ? Oct 23, 2018 17:26 |
|
It's also probably possible to specify how many old kernels to keep whenever apt goes through the removal process. I know it's possible in Fedora because I've got a small VM set up, to use the current working kernel and keep only the previous one spare. I'm on my phone right now but the method for Ubuntu should be just a Google away.
|
# ? Oct 23, 2018 17:37 |
|
VikingofRock posted:What makes Solus a special snowflake distro? I don't really know much about it other than the fact that it's rolling release and people tend to like it. Hey, I like it too! They have some good, fresh ideas but that doesn't make them super easy to use with other OSes. See the bootloader section here: https://getsol.us/2018/02/15/exploring-solus-architecture/
|
# ? Oct 23, 2018 17:53 |
|
Horse Clocks posted:The + register is also your system clipboard.
|
# ? Oct 23, 2018 19:00 |
|
I installed vim 8.1 from Homebrew and it yanks to/pastes from the system clipboard by default and I can't figure out how to turn it off
|
# ? Oct 23, 2018 19:36 |
|
touch ~/.vimrc
|
# ? Oct 23, 2018 19:49 |
|
jaxercracks posted:So you can see that the current kernel stuff takes up right at 70mb. Which gives you just barely enough space to have three kernels in there. The kernel autoremove script indicates that depending on how it is called you might end up with up to four kernels. So it seems like enough space for four kernels plus a bit of slack would be the correct amount. You might be able to reduce the size of the initrd.img by editing /etc/initramfs-tools/initramfs.conf . Set MODULES to "dep" and COMPRESS to something other than "gzip". Admittedly I haven't tried these myself. 'man initramfs.conf'
|
# ? Oct 23, 2018 20:11 |
|
For how long has Centos had security errata in yum (ie yum --security update)? I thought that was a RHEL only thing and was surprised to see it work the other day.
|
# ? Oct 23, 2018 21:20 |
|
|
# ? Apr 20, 2024 03:40 |
|
my bitter bi rival posted:For how long has Centos had security errata in yum (ie yum --security update)? I thought that was a RHEL only thing and was surprised to see it work the other day.
|
# ? Oct 24, 2018 03:06 |