[09:34] <dholbach> hiya
[09:35] <dholbach> a colleague of mine is running into https://bugs.launchpad.net/cloud-images/+bug/1792543 - who could help out with this?
[09:42] <vlt> Hello. In a (typical?) RAID10 => LUKS => LVM setup, where would you insert a fast SSD (or two of them as RAID1) as bcache device? And does the RAID10 then still make much sense or would you replace it with RAID5 or 6?
[09:54] <dholbach> Odd_Bloke: ^ can you help with this bug or is it somebody else?
[11:49] <openfire> vlt: Frankly, I would never replace RAID10 with any kind of parity RAID setup. As for the caching option...
[11:49] <openfire> vlt: I would add the SSD as an LVM caching volume.
[11:51] <openfire> vlt: Rather than using bcache/dm-cache directly.
[11:52] <openfire> vlt: man 7 lvmcache
[12:23] <cryptodan> this weekend my dell pe 4600 will be retired and replaced with a more modern hp mp350l dual 2.0ghz hexacore setup
[12:35] <vlt> openfire: Thanks for your answers. I'll definitely have a look at lvmcache. Why would you never use parity RAID instead of RAID10? I thought parity RAID's weakness is writing and that should be covered by the caching device. What other reasons are tehre?
[12:44] <openfire> vlt: Undetectable data corruption due to read errors during rebuilds. Most (all?) storage vendors consider parity RAID a dinosaur these days.
[12:46] <vlt> Ah, ok.
[15:00] <frickler> jamespage: can you please check https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1750848 ? I'm confused how a rebuild of mimic can fix an issue with luminous.
[15:01] <frickler> nice name btw, cosmic disco ;)
[15:02] <jamespage> frickler: some of the changelog history is missing from the comment; I did a load of work to cleanup the lintian awarnings on the mimic packages
[15:02] <jamespage> I won't backport that
[15:02] <jamespage> it does not materially effect the function of the package
[15:03] <frickler> jamespage: o.k., understood, but the plan would still be to support latest 12.2.x for the time being, just not fix that issue there, correct?
[15:03] <jamespage> yup
[15:04] <frickler> ok, thx
[15:23] <dholbach> hiya jamespage - how are you?
[15:23] <dholbach> do you know who could help with https://bugs.launchpad.net/cloud-images/+bug/1792543?
[16:09] <zetheroo> I am trying to setup static IP on 18.04 server by editing /etc/netplan/50-cloud-init.yaml
[16:10] <zetheroo> So far I can ping internal IP's and external domain names (google.com) but I cannot ping any internal hostname's
[16:11] <zetheroo> this is what the config looks like https://paste.ubuntu.com/p/rshmvDyC5X/
[16:12] <zetheroo> what am I missing?
[16:12] <teward> zetheroo: does it give you a specific error message?  have you verified they're all on the same subnet?  Are they on different subnets and there's a firewall restricting access?
[16:12] <teward> zetheroo: is there an internal search domain?
[16:12] <teward> check another system for any 'searhc domains' that it defaults to
[16:12] <zetheroo> teward: I am trying to reach hostnames on the same network in the same domain
[16:12] <teward> that might be necessary
[16:12] <teward> for the DNS resolutoin to work proper
[16:13] <zetheroo> yes, but where do I enter that info?
[16:13] <teward> https://netplan.io/examples#dhcp-and-static-addressing
[16:13] <teward> nameservers > search
[16:13] <teward> underneath the ethernet device config
[16:13] <zetheroo> ah ok
[16:14] <zetheroo> brilliant - thanks
[16:59] <cyphermox> teward: thanks
[17:09] <dougquaid> I'm in the process of hardening an ubuntu server. Is there some software that I can scan the server with that will find vulnerabilities?
[17:10] <lotuspsychje> dougquaid: i like lynis as nice scanning tool, but security has very wide ranges
[17:10] <lotuspsychje> dougquaid: you can harden your server on so many different ways
[17:10] <lotuspsychje> dougquaid: first thing is gathering info on everything, like a hacker would do
[17:11] <lotuspsychje> !security | dougquaid
[17:11] <dougquaid> thanks
[17:12] <lotuspsychje> dougquaid: nmap is also very good for information gathering on your ip's/locally and remotely
[17:13] <lotuspsychje> dougquaid: can you share with us wich services you run from your server?
[17:14] <teward> cyphermox: for?  oh, the netplan thing with zetheroo?
[17:20] <giaco> hello
[17:22] <giaco> I am installing ubuntu server on a machine that has no graphical output, but I have a serial terminal 115200n8. I am supposed to pass additional kernel command line parameters such as "console=tty0 console=ttyS0,115200n8" but I can't find in the install menu where I am supposed to type it
[17:24] <giaco> By sending F3 it says "To use one of these boot methods, type it at the prompt, optionally followed by any boot parameters. For example: boot: install acpi=off", but if I do "install console=tty0 console=ttyS0,115200n8" it says "Loading install... failed: No such file or directory"
[17:25] <giaco> same for "expert console=tty0 console=ttyS0,115200n8" "Loading expert... failed: No such file or directory"
[17:25] <cyphermox> teward: yup
[17:27] <giaco> but if I do "memtest" it goes "Loading /install/mt86plus... ok", so the boot prompt seems working. Are "install" and "expert" correct references?
[17:29] <teward> cyphermox: ah, cool.  yeah i keep that examples page bookmarked :P
[17:29] <teward> since it explains some of those questions real easy :)
[17:29] <cyphermox> teward: yup, and I'll do one big update of it very soon
[17:29] <giaco> could you please confirm me that the "install" and "expert" boot options are correct? The documentation is mesleading
[17:31] <teward> cyphermox: cool, let me know when that happens, can't wait to see even more examples and know even more about Netplan!  :D
[17:37] <giaco> thank you for your consideration. I'm dropping ubuntu server for a good debian, which has a working install bootloader
[19:01] <Ussat> wat ?
[19:01] <Ussat> I have never had an issue with a Ubuntu install bootloader
[19:02] <mason> Ussat: You need to be more special to encounter it.
[19:02] <Ussat> I see, reading up
[19:03] <Ussat> I have not installed a debian in a long time, but I dont expect its a lot different
[19:04] <mason> Not for the server install, certainly.
[20:13] <xnox> giaco, i typically use F6 to modify the cmdline of the installer.
[20:14] <xnox> giaco, i don't think failing to pass the right cmdlines is considered to be a broken bootloader, especially since this is installer only and not the installed system.
[20:15] <giaco> xnox: contrary to what you are saying, failing to add kernel parameter at boot, especially at intaller boot, can be considered a severe fault
[20:16] <giaco> xnox: moreover, using the ubuntu netinstall image the behaviour is different and the same command line options are working
[20:16] <xnox> giaco, are you using non-ubuntu netinstall images and those are broken?
[20:16] <giaco> so is the the ubuntu server 18.04 iso (the ~900MB one) that is not working
[20:17] <giaco> xnox: no, the ubuntu netinstall is working, the normal iso is not
[20:17] <xnox> giaco, live or non-live one? cause the live one, which uses subiquity, has probably incorrect help text on the scren.
[20:17] <xnox> giaco, right, we should fix that. One second, let me give you two links.
[20:18] <xnox> giaco, http://cdimage.ubuntu.com/releases/bionic/release/ubuntu-18.04.1-server-amd64.iso  this is the "old" server installer, which uses d-i, and should use all the same kernel cmdline parameteres as netinstall image does.
[20:18] <giaco> this is the one that did not work https://www.ubuntu.com/download/server
[20:19] <giaco> xnox: I don't think is a matter of bootloader feature availability: the parameters are passed straight to the kernel
[20:20] <giaco> btw I've succeeded in installing ubuntu server 18.04 using netinstall + parameters "console=tty0 console=ttyS0,115200n8"
[20:20] <xnox> giaco, eeeee..... well..... some of them are also parsed by initramfs, and also parsed by installer, to launch the installer on the right console.
[20:21] <xnox> giaco, i suspect that the one pointed at https://www.ubuntu.com/download/server is a subiquity based one, and doesn't in fact support multiple consoles and doesn't start installer on ttyS0
[20:21] <xnox> only starts it on tty0
[20:21] <xnox> actually tty1
[20:21] <giaco> well, I need ttyS0 anyway, which is not started by default on netinstall
[20:22] <xnox> giaco, we have two server .iso; this looks like a legitimate bug in one of them; please try this iso http://cdimage.ubuntu.com/releases/bionic/release/ubuntu-18.04.1-server-amd64.iso
[20:22] <xnox> this one must support console=ttyS0,115200n et.al. correctly; and the same params as netinstall does, as they are built similarly.
[20:22] <xnox> the image from https://www.ubuntu.com/download/server is very different beast.
[20:22] <giaco> I didn't know that, sadly
[20:22] <xnox> giaco, or like for other architectures look at http://cdimage.ubuntu.com/releases/bionic/release/ to download for the whatever arch you need it for
[20:23] <giaco> But now I'm into configuring the new system, I don't really have time to go back and retry
[20:23] <xnox> giaco, i'm sorry =/ this is quite low-level details, and i know this because well, i build these images.
[20:23] <giaco> Can I keep it as homework for next time?
[20:24] <xnox> giaco, i will open the bug report to fix console=ttyS0 support on the other image (the one from https://www.ubuntu.com/download/server )
[20:24] <giaco> xnox: oh, well. Thank you for helping me
[20:24] <xnox> giaco, btw. on the https://www.ubuntu.com/download/server there is a link to these other images
[20:24] <giaco> I'm curious: why splitting?
[20:24] <xnox> giaco, under the green button tehre is "see our alternative downloads." in blue
[20:25] <giaco> is there a "dev/prod" into ubuntu server installer ?
[20:25] <xnox> giaco, we are working on making the new one; but clearly it would be irresponsible to stop building the old one, given that it is impossible to find all corner cases.
[20:25] <giaco> xnox: right. Well, glad to help
[20:25] <xnox> cause we re-wrote the installer from scratch (the new one)
[20:26] <xnox> plus we dont' have encryption yet in the new one
[20:26] <xnox> giaco, so yeah, hence the download page does say "This release uses our new installer, Subiquity. If you need support for options not implemented in Subiquity, such as encrypted filesystem support, the traditional installer can be found on the alternative downloads page."
[20:26] <xnox> or in your case "has buggy console="
[20:26] <xnox> =)
[20:28] <giaco> xnox: well, you know, I just went through XD
[20:28] <xnox> giaco, can you explain again to me your system you are using? you need to boot it using console=ttyS0,115200n8 but it is otherwise x86_64 machine right?
[20:28] <xnox> with a serial connection / bootloader on the serial console?
[20:28] <xnox> https://bugs.launchpad.net/subiquity/+bug/1770962
[20:28] <giaco> xnox: yes, this https://www.pcengines.ch/apu3a2.htm
[20:29] <xnox> thanks.
[20:29] <xnox> giaco, for now please use the alternative iso, i pointed you to =/
[20:29] <xnox> also, once installed, Ubuntu is so much nicer than Debian ;-)
[20:30] <giaco> xnox: and I connect with this https://www.pcengines.ch/usbcom1a.htm
[20:30] <teward> xnox: subiquity supports RAID (software, FAKERAID, hardware RAID controllers-controlled arrays, etc.), right?
[20:31] <xnox> giaco, thanks.
[20:31] <xnox> giaco, and when you booted that broken iso; it did boot and show bootloader (cause you are describing things like F3 and etc) but you didn't manage to boot to installer, correct?
[20:32] <giaco> thanks to you. I didn't know I was looking at something new. I tend to consider the boot part of solid server OS quite bulletproof. I not even considerer an "old option". My fault
[20:32] <xnox> giaco, this is helpful. i'm sorry you had this experience, but i'm glad you shared it with us.
[20:32] <xnox> giaco, i might want to update the website to also mention that serial console is not working.
[20:33] <xnox> giaco, note, on all other architectures we are still pointing to old installer, as those are typically done via non-graphical terminals.
[20:33] <giaco> I confirm. I was able to see the bootloader menu, preff F-keys, type at boot: prompt, start memtest. But install, expert, cli and other were not accepting console options. Without them I was that it was loading something but hey no feedback no party
[20:33] <xnox> Cool, thanks!
[20:34] <giaco> * Without them I saw that something was loading
[20:34] <xnox> giaco, yeah, most likely there was a very pretty installer on tty1 which you couldn't see.
[20:35] <xnox> giaco, if you launch the new ISO in a qemu/kvm you will see how simple, and pretty it is. Full-terminal curses ui, with minimal steps, import of ssh keys, and very quick install speed.
[20:36] <giaco> I tend to be the guy with null-modem cables in 2018
[20:36] <xnox> =))))))))))))))) but also! thank you =)
[20:36] <xnox> giaco, i might try to fix this, and will try to ping you when it's done.
[20:36] <xnox> giaco, can you subscribe to https://bugs.launchpad.net/subiquity/+bug/1770962 ?
[20:37] <xnox> just added a comment to that bug report, about your experience.
[20:37] <xnox> teward, LVM, RAID, Bonds, Vlans => are all there.
[20:38] <xnox> teward, fakeraid is partial
[21:10] <giaco> ok, new network manager. What is netplan? And why is not picking my interfaces? Error in network definition //etc/netplan/01-netcfg.yaml line 9 column 10: br0: interface enp1s0 is not defined
[21:10] <giaco> but ip a | grep enp1s0: enp1s0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000