« first day (2469 days earlier)      last day (2479 days later) » 

12:47 PM
@derobert You asked for a screenshot earlier. See ^^
As you can see, no RAID or LVM option.
@derobert Never mind, my error. Looks like I had to delete the filesystem to get those other options to show up.
 
 
1 hour later…
2:01 PM
@Pierre.Vriens just in case you can see this: thank you for your edits/improvements! Something that I learned when starting to edit, realize that if you edit the text a closed post it puts it into the ReOpen queue, so please continue to write good comments about what you changed, so that the reviewers know it was just typo-fixes and not an attempt to re-open the question. Thank you!
 
@JeffSchaller He can't. Only people who've been in the room in the past few days (2 weeks?) can be pinged.
But if he's edited a post, you can leave him a comment there and he will be notified although he won't show up in the autocomplete.
 
2:21 PM
@terdon cue the sad Game Over sound of me not knowing one more thing about the site, but cheers for the excellent work-around idea -- will do; thank you!
 
@JeffSchaller Snort, I get that sound quite often myself and I've been here longer.
 
Hmm, am I just wrong? I thought edits to closed Qs put them into the reopen queue, so either I'm wrong or I'm not seeing them.
 
2:43 PM
@JeffSchaller They should, yes. Link?
But you won't see it if you've already reviewed it.
 
@terdon I reviewed the suggested edit (and approved it/them), for example: unix.stackexchange.com/review/suggested-edits/201071 -- but then expected to see items in the Reopen queue @ unix.stackexchange.com/review/reopen/history
ok, brain fart on my part for that one: " It needs more reviews from other users to be completed"
might be a better example
 
Ah, I don't think editing a dupe will put it in the review queue. Not sure though.
 
oh boy, there's an exception I won't remember :) No big deal; I left a comment as you suggested, and pierre was receptive. Thanks for taking a look!
 
@JeffSchaller No, I don't think it is an exception actually. Looks like there might be a time limit though:
19
A: Which edits push closed questions to the reopen review queue?

Shog9From a time when I probably knew something about this: Currently, this means that a closed question will automatically be added to the reopen queue when it is... ...Edited (body edits only) within 5 days of closure by the author. Or, ...Edited (body edits only) within 5 days of clos...

 
3:01 PM
@terdon whew, good find - thank you! I also don't stand a chance of remembering that, either, but it seems to be a sane approach.
I must have stumbled across the "inadvertent reopen queue" situation when I was editing recently-closed Qs
I should have known, there's a Q&A [here](meta.unix.stackexchange.com) or [here](meta.stackexchange.com) or [here](meta.stackoverflow.com) for almost everything
well those links didn't work sigh stopping for a while :)
 
@derobert Try the following: put a filesystem on an LV (ext4 will do), and then feed it to a libvirt VM running the Debian installer. I think you'll see that the choices it shows are quite limited, and don't include RAID or LVM. And as for as formatting a particular device, it refuses to include RAID or LVM even if you delete the filesystem partition. I know this sounds weird. I could try to come up with a repro recipe.
I also accidentally downloaded a i386 installer DVD instead of an amd64 one. Go me.
 
3:19 PM
when you install a module in /lib/modules/$(uname -r)/kernel/drivers, is it auto-detected by depmod or is there some kind of configuration file that has to be modified?
 
Hello, I think I need help in connecting with ssh over socks5
I'm using this string:
ssh -o ProxyCommand='nc -x myproxyserver.example.com:1080 %h %p' \
targetsshserver.example.com
However it seems my nc hasn't the -x support and no proxy support also
 
what device?
you're likely using netcat-traditional instead of netcat-openbsd, which has newer features like proxy
14
Q: What are the differences between netcat-traditional and netcat-openbsd?

spezticleWhat (if any) are the significant differences between netcat-traditional and netcat-openbsd? I'm having trouble finding any relative information. Anybody familiar that can offer some insight?

 
3:36 PM
Oh so i have to apt-get install netcat-openbsd ?
I'm on debian9
 
4:20 PM
@FaheemMitha errr... ok... I'll try quickly.
ok, so used virt-manager to create a new vm, and a new lv (vm_wtf_1). About to hit begin install. Just ran mkfs.ext4 /dev/mapper/Zia-vm_wtf_1 on the host.
Using text-mode installer with Stretch DVD 1. Normal install, not expert.
Partition disks: manual. I see what you mean....
Why did your fresh LV have a ext4 volume on it? Anyway, found a workaround in the installer: select the ext4 volume, tell it to erase it. You can cancel a little way through. Then from the partition disks menu, select "go back". Then select "paritition disks again"
Then you can select vda, it'll let you create a new partition table.
 
4:38 PM
Hi!
 
@user3450548 yup, that's right. install netcat-openbsd then use sudo update-alternatives --config nc to switch to it
 
I had some problems with the ssh server on one of my computers and decided to reinstall it. I installed both openssh-server and openssh-client. The installation seemed succesful, but I have no ssh_config only the sshd_config file. Does this mean something went wrong?
 
@ÉtienneBézout ssh_config should come with the client. Which distro?
 
5:03 PM
Sorry, had to go afk. It's debian.
I purged ssh and openssh-server.
And then reinstalled.
Also manually deleted /etc/ssh/
 
@ÉtienneBézout You needed to purge openssh-client as well
Otherwise Debian is going to treat the deleted config file as an admin change that must be preserved.
BTW: if anyone has a CentOS 7.2 box to check:
0
A: Accidentally erased contents of /etc/fstab in Centos 7.2

derobertMaking this community wiki to invite contribution from someone familiar with CentOS/RHEL 7.2 The UUIDs come from your blkid output. The paths come from the other output, and the filesystem type and options from /etc/mtab. The dump and fsck order fields are guesses. (I used the same fsck pass bec...

 
What's the difference between the ssh package and openssh-client?
 
The description of the ssh package explains:
This metapackage is a convenient way to install both the OpenSSH client
and the OpenSSH server. It provides nothing in and of itself, so you
may remove it if nothing depends on it.
 
Also, is it ok if I manually put the config file there myself? I have one from another install.
 
@ÉtienneBézout sure. That's fine.
 
5:12 PM
Ah, ok thanks!
 
You could also recover it from the package directly using dpkg-deb -x (to some random directory, then copy the file out of there)
 
Ok, thank you very much for your help!
 
@ÉtienneBézout once you've fixed that, please apt-get install etckeeper. That will keep your /etc in version control (git by default), so you can easily recover from mistaken configs.
 
Ok, will do. Thanks for the tip!
I have to go and eat now. I'll fix it later and drop a comment to confirm that I got it right.
Bye!
 
5:49 PM
@derobert Sneaky. But there should be a standard option that lets you wipe the partiition table. Can you think of a reason they didn't provide it?
 
@FaheemMitha Normally selecting vda will do that... but with a filesystem on the whole device, it did nothing. Not sure if that's a bug or intended.
 
Another usability issue - the installer seems to make it as hard as possible to run grub-install /dev/sda as well as /dev/sdb. You just get one device. Unless I'm missing something.
@derobert I vote for usability screwup. Worth reporting?
 
@FaheemMitha errr... it should ask you about that. And /dev/vda would be the right device (unless you switched the disk to SCSI)
@FaheemMitha I guess.
 
@derobert Yes, I switched to scsi. Though should I do that in future?
@derobert You don't sound convinced.
 
@FaheemMitha Then it should offer to install grub on /dev/sda for you. Or possibly just do it.
 
5:53 PM
Also, I noticed that if you select RAID 1, choose 2 devices, but only select one of them, it's happy to do so. Shouldn't it abort?
@derobert Sure, but it won't let you install on both. I tried quite hard just now.
Anyway, installation is a success! Yay.
 
@FaheemMitha right, yeah, if you want /dev/sdb as well, that's harder... I don't think the installer offers that (other than execute a shell, or switch to a different vt). Of course, you can do it once the system is booted.
@FaheemMitha No, perfectly valid to create a degraded raid1.
 
@derobert I dropped to the shell. ash said it didn't know what grub-install was. Not exactly helpful.
 
@FaheemMitha chroot /target then try again. But anyway, just do it after reboot
 
@derobert Oh? Shouldn't it at least give a warning? Seeing as you just selected 2 devices?
 
@FaheemMitha I don't know. It's reasonably common to do it, as it makes the install faster (since the install won't be competing with the initial sync, or writing twice).
 
5:56 PM
And to answer your question, I tend to create a filesystem on new LVs. Just force of habit, I guess.
@derobert Oh. So not worth reporting?
 
@FaheemMitha I guess you could. If there were a way to give a non-intrusive warning, I suppose it's nice polish.
 
Btw, just got an error/warning on boot with the VM (I see this with my actual system as well):
> Failed to connect to lvmetad. Falling back on device scanning. But it still boots.
 
Yeah.... I see that too a bunch. Haven't bothered to track it down. Maybe it comes from initramfs. Not sure.
 
Is there a keyboard command from switching between full screen and back? I thought I saw F11, but maybe I'm confused. Plus F11 isn't doing anything anyway.
@derobert In view of my LVM clone debacle I was curious.
Which I'll try again on the VM. Maybe I should recite "The Charge of the Light Brigade" while I do it.
 
@FaheemMitha I do suggest grabbing a acoustic guitar and setting it as a folk ballad, too.
 
6:03 PM
@derobert I used to play the guitar as a child. I'm out of practice.
 
Well, does it matter when it's a folk ballad?
 
Plus TCOTLB isn't terribly singable.
@derobert I don't know. I'm not even sure what a folk ballad is.
Like "Scarborough Fair"?
 
Yeah, that's probably one.
 
That's a nice song.
I'm thinking of the Simon and Garfunkel version, of course.
So, is there any way to wipe the partition table without cheating?
 
@FaheemMitha without cheating? I think selecting and picking "wipe" is pretty non-cheaty...
Of course, when you create the LV, by default LVM wipes the first few blocks for you...
 
6:07 PM
@derobert Hmm, you wrote:
2 hours ago, by derobert
Why did your fresh LV have a ext4 volume on it? Anyway, found a workaround in the installer: select the ext4 volume, tell it to erase it. You can cancel a little way through. Then from the partition disks menu, select "go back". Then select "paritition disks again"
As far as I can tell, the "cancel a little way through" is actually necessary.
 
yep, I don't know of an easier way. Other than switch to vt 2 and use dd
@FaheemMitha It shouldn't be—the "go back" part probably is though
I just got sick of waiting, so I canceled part way through
 
If you let it erase the filesystem, it doesn't seem to work.
 
Even with the go back step?
 
@derobert Ah. I'll have to go back and try again.
@derobert Yes, I tried a bunch of stuff. I'm nothing if not persistent.
@derobert Does ash know about dd?
 
Yeah. The go back and select partition disks is essential. Otherwise the installer doesn't notice the change
@FaheemMitha I'm pretty sure there is a dd in the installer environment. Probably busybox's dd
 
6:10 PM
@derobert Hmm, not sure if I went back to partition disks.
 
You go back to the main menu. Then select partitions disks from that menu.
 
@derobert Ok. Still, you'd have to figure out the correct syntax, with your only machine temporarily out of commission.
@derobert I'll try setting up another VM and give it a try.
 
@FaheemMitha well, depends on if dd --help does anything useful
 
@derobert So you don't think the erase part is necessary?
 
but dd if=/dev/zero of=/dev/vda isn't that bad... Of course, cp /dev/zero /dev/vda would work too. Or cat /dev/zero > /dev/vda.
 
6:13 PM
@derobert Don't you just need to wipe the MBR?
Though I didn't check to see if a GPT installation was possible.
 
@FaheemMitha I think it needs to erase the first few blocks, to wipe the ext4 signature from the block device. Then you need to exit the partitioning menu (by going back to the main menu). After that, when you re-pick partition, it'll rescan the disks and not find the filesystem there.
@FaheemMitha it is, if you use expert mode, or if you go to the main menu and change the debconf priority. Then it'll ask you what type of partition table to create—I think you a dozen choices, msdos and gpt are some of them.
 
@derobert I see. That makes sense, but isn't immediately obvious. Wouldn't it be better to just have an option to scrub the partition table?
@derobert I didn't see GPT. And this time I didn't try expert mode. So many options, so little time...
 
@FaheemMitha There is supposed to be — hit enter on the disk line (not one of the partition ones). It should ask about creating a new empty partition table. But it didn't. That could be a bug.
 
@derobert It doesn't do that here.
It just returns.
And yes, that's what one would intuitively expect.
Anyway, probably worth sending a "random comments about the installer" post, for the installer team to ignore.
 
yeah, it does that if there isn't a filesystem there. But when there is a fs, then it just does nothing. That seems like a bug. It should at least tell you why its refusing.
 
6:19 PM
@derobert Ok. Thanks for the confirmation.
I think the customary thing is to post against the virtual "debian-installer" package, or some such.
@derobert BTW, you write:
> I don't have a CentOS 7.2 machine to check.
I'm surprised you don't have a virtual CentOS machine handy...
 
Didn't feel like downloading DVDs and installing one. Since someone else surely has one running!
 
@derobert Sure. But I thought you'd have one handy already.
 
Nope. Don't run CentOS here. Only Debian.
 
7:21 PM
This is an odd message:
(Which it seems I am unable to paste. Bummer.)
This seems overkill just to get copy and paste to work from inside a VM - unix.stackexchange.com/questions/109117/…
 
8:21 PM
@FaheemMitha hey, there's a reason I keep using screenshots
normally I just ssh in to the VM... then it's easy
They use SPICE by default now, btw.
 
9:00 PM
@derobert Ok, I got it to work. Thanks!
 

« first day (2469 days earlier)      last day (2479 days later) »