[00:35] i need help!!!! [00:36] anyone professional with ubuntu server ? i'm a student and learning [00:51] papajo: you should just ask your question(s) [01:21] sorry i was distracted , i'm trying to set up my cyberduck ftp client to be able to upload to my www folder on my ubunutu server and i'm not sure what i'm doing wrong, i gave all the correct permission for my user i followed the ubunutu server documentation [01:24] i also tried Configure the user's home directory and i get this usermod: user ************ is currently used by process 1759 [01:50] From usermod manpage. "You must make certain that the named user is not executing any processes when this command is being executed if the user's numerical user ID, the user's name, or the user's home directory is being changed." [01:52] This why it's advantagous to have a second user with sudo ability, in case the primary one requires doing something like this with usermod === nchambers is now known as uplime === chat is now known as Guest64523 [03:25] Figure I'll ask here as well as the #ubuntu channel - I'm trying to determine the root cause for a consistent system hang issue I'm running into. At random, during approx 8-48 hour periods, the server hangs to a black screen. It's not recoverable by killing X, by switching virtual terminals, or with REISUB. I have to physically reset the machine. [03:26] Today the system hung and there was something printed to the console: https://i.imgur.com/wzl5HuI.jpg [03:27] Any ideas appreciated. [03:32] stiltzkin: that looks like it died on boot, as it couldn't get a root device. Look at regenerating initramfs, updating grub, etc. [03:38] But the system boots correctly [03:38] upon reset [03:38] so I have no idea why that would be the case [03:41] Oh, that's a bit odd. [03:42] stiltzkin: Maybe check SMART status or somesuch. Maybe you have a marginal disk. There's probably a better explanation, but I'm not sure what it would be. [03:42] SMART checks out [03:42] the message printed to console confuses me since it should not be looking to /dev/md* for the root disk [03:42] stiltzkin: So, the system couldn't find a component of its root MD-RAID1 or somesuch... Maybe it didn't spin up on cold boot, in time, but did on reboot? [03:43] Root is not on the RAID [03:43] this is a fileserver with a single host SSD [03:43] Oh! Heh. [03:43] so it shouldn't have to look to /dev/md* for root [03:43] Welcome to systemd making everything so very nearly transparent. [03:43] Yeah. [03:43] lol. [03:45] I have noticed one other tidbit: sometimes I will recover from one of these hangs with a hard reset and the system will not boot because the UEFI drive to boot from in BIOS becomes randomly unchecked. I have to boot into BIOS, select 'ubuntu' as the UEFI drive to boot from, and reboot to get the system up [03:45] Also a bit odd. [03:45] That to me makes me think the drive is failing (responding too slow on boot to show as a boot target?) but the drive seems fine [03:46] Lately dpkg-reconfigure grub-uefi (IIRC) erases my boot entries entirely and errors out, which is fun on kernel upgrades... [03:46] hmm so maybe that's something else then [03:46] Yeah, that's a bit odd too. Maybe see if there's a firmware upgrade available...? That seems like it should be related, that UEFI bit. [03:47] The funny bit is, with a single disk... where did it get the initramfs and kernel, if the root disk isn't available? [03:47] for the drive? there may be. it's a 2 year old SSD [03:47] If you can get it back into busybox sometime, poke around and see what devices are there, etc. [03:48] in the picture I sent the system is completely hung so there's no chance of keyboard input [03:48] but I have it fully back up now, everything seems normal [03:48] Ah... That's also not a fantastic sign. [03:48] But nothing is leaping out at me as a plausible cause. [03:48] but there's a good chance when I check in the morning, the system will be down. Or if not, then it will be down sometime tomorrow afternoon or tomorrow night. Guaranteed [03:49] So, it's spontaneously rebooting to get you to that spot? [03:49] That seems like marginal hardware. [03:49] I've actually never seen the screen I sent before. Every other time it does this it hangs to a completely black screen [03:49] First time I've seen those messages is today [03:50] Either way... A spontaneous reboot is a bad sign, and I'd be hard-pressed to blame even systemd. [03:50] running extended smart test on the host ssd now... [03:50] I'd also maybe watch heat, watch dmesg remotely... Let's see... [03:51] stiltzkin: https://www.cyberciti.biz/tips/linux-netconsole-log-management-tutorial.html [03:51] Get that going, and maybe you'll see badness emerging before whatever event is taking the system out. [03:51] yeah that's a good idea [03:52] yeah the test checks out, looks normal on the drive [03:52] Sometimes if you have an issue that freezes you up like that, the system won't log anything, you won't be able to take a crash dump, nothing, but the udp log can catch the initial badness. [03:52] mhm [03:52] good advice [03:53] I had to use that recently to confirm that Nouveau is bad and doesn't work on my work laptop. [03:53] this sounds like a job for the raspberry pi in my drawer [03:53] That could do it. [03:55] Note that I've had issues with predictably bad interface naming and netconsole. [03:55] If you have issues setting it up, try net.ifnames=0 on your kernel command line. [03:56] OK [03:56] Good to know, thank you [06:35] hallyn: i may well have talked of one before. [06:35] i think there might be one in buildroot...or in yocto [06:36] i think i'd just use libguestfs or a qemu in some way and do it all in the guest. heavier but more flexible. [06:37] yeah libguestfs is an option, as is us fuseext2. (using fuseext2 right now but it hangs on umount) i'll look at buildroot, thx [06:42] oh, maybe i can use e2tools? [06:47] yup that'll work :) [06:57] Good morning [12:19] jamespage: coreycb i can see that zunclient is packaged, is there any plans on packaging the zun project? [12:41] zul: no immediate plans that i know of [12:42] zul: would be happy to help usher it through if it were contributed of course [12:42] zul: sorry! [12:42] tobasco: ^ [12:43] * coreycb needs coffee [12:43] coreycb: Freudian slip? [12:45] zul: possibly :) [12:45] heh [12:45] * zul goes back to enjoying caffeinated beverage [13:09] coreycb: ok thanks for the info, sorry to bother but is there any docs on the packaging procedure for a packaging-novice? :) [13:15] tobasco: sure, i'll paste a couple of links [13:15] for working with Ubuntu OpenStack packages in general: https://wiki.ubuntu.com/OpenStack/CorePackages [13:16] Ubuntu packaging overview: http://packaging.ubuntu.com [13:17] tobasco: for a new package i would typically start by cloning the source for say, keystone, as a starting point [13:18] and then start updating it for the new project [13:20] tobasco: package source is all here btw: https://code.launchpad.net/~ubuntu-server-dev/+git [13:26] coreycb: thanks! i will educate myself and hopefully learn a few things :) [13:28] tobasco: cool, i'll be here if you have questions. [13:40] is 18.04 frozen now, there an installation image somewhere to test all the interesting ways my software will break on it before it's out? [13:41] jelly: Feature-freeze last week but a lot can still change as bugs get ironed out [13:42] of course [13:43] this might be a good starting point http://cdimage.ubuntu.com/daily-live/current/ [13:44] well! that saved me some copy/pasting :p [13:44] thanks [13:44] or the daily/ (as opposed to daily-live) [13:44] I wish we had a nice guide to all the various images [13:44] even if I'm going to prefer http://cdimage.ubuntu.com/ubuntu-server/daily/current/ and try to plug it to our pxe/preseed thingy [13:45] there's something like thousands of them and my firefox history doesn't have them all :) [16:18] ahasenack: https://code.launchpad.net/~nacc/usd-importer/+git/usd-importer/+merge/336804 [16:18] rbasak: were you able to look at my pushes to add tests to the importer? [16:22] nacc: yeah. I'm still pondering that. [16:23] rbasak: ack, some of the fixlets in that branch are probably not ideal -- I tried to document in the MP what I wanted to test for [16:23] rbasak: I think I'm correctly testing for what I described; perhaps we can extend the tests in the future, if you have further criteria? [16:25] nacc: it's the structure of the tests themselves. I think I'd like to see what you defined what you wanted to test represented as a data structure of your expectations, rather than checking each individual thing in code. However, I don't think it's needed to block at all, in that if you're testing what we want to test, it's fine to land and then refactor the tests later. [16:25] nacc: but I think that branch is still a wip, right? [16:25] rbasak: do you have an example (even pseudocode) of what you mean by a structure of expectations? [16:26] rbasak: you mean like something that defines expected import tags, expected orphan tags, expected branch commits, etc? [16:26] rbasak: https://code.launchpad.net/~nacc/usd-importer/+git/usd-importer/+merge/340874 this one is 'ready' [16:26] rbasak: the other, which actually includes real fixes is WIP, as I need to write the tests I mentioned in that MP [16:27] nacc: it'll be a really common pattern for importer tests I think. "When the set of publishes looks like X, I expect to see the imported git structure to look like Y" [16:27] rbasak: yeah, that's a good point [16:28] rbasak: just trying to figure out how to descirbe the git structure without having to describe hashes, etc. [16:28] nacc: separately, I think I'd like to try to more formally specify the importer in terms of the results. [16:28] nacc: yeah, that's the challenge. [16:28] rbasak: unless what i wrote is what you meant (tags, branches, etc) [16:28] nacc: combining the two things, it'd be nice if the specification then maps directly to the tests. [16:30] nacc: got time for a HO? [16:30] rbasak: ack [16:30] rbasak: standup? [16:30] omw === tobasco is now known as tobasco|away [16:59] Hello. I'm having an issue. I have ubuntu 16.04 installed and its been running fine for a while. However now when I try to access the server via \\ip i keep getting not found [17:00] DexDeadly: I don't entirely follow; what server are you trying to access, and how are you trying to access it? [17:00] I have samba installed on the ubuntu server but I can't access it [17:00] it used to work but it is not any more [17:00] Did you make any changes? [17:01] I can ping it [17:01] I didn't. One thing I noticed recently was my /dev/sda1 boot was full so I cleared that, made sure all the apt-get update/upgrade all was done [17:01] DexDeadly: Is the Samba service definitely running? [17:02] sudo service status smbd says active [17:03] I've never really used Samba, I'm afraid, so I'm now out of my depth. [17:05] ok [17:05] DexDeadly: check there are no firewall rules blocking it; check the various samba services are listening on the correct network interfaces with "sudo ss -tnlp" [17:07] hmm ok i just disabled my windows firewall, i thought it was [17:08] now im getting can not access === uplime is now known as nchambers [17:53] rbasak: 100% phasing is 4943 packages [18:30] Hey all. How do I use /etc/cron.d ? I have a file called "test" that has a cronjob in it to call a script but it does't work? Do I have to enable it or something? [18:53] So I still can't seem to get in, anything else? [19:43] Is my smb.conf correct. This worked before [19:43] https://pastebin.com/BGN9an0i [19:44] I still can't access the samba share [21:03] sorry lost connection. Anyone have an idea of what else I can try [21:09] your smb.conf is 100% wrong [21:09] atleast if you want to connect to anything >windows xp [21:14] may I ask what [21:14] DexDeadly: it may be easier for others to review your samba conf if you pasted the output of "testparm -s" instead of the smb.conf file [21:14] cause it used to work [21:14] sure i can do that [21:14] define, used to work [21:14] something changed [21:14] like upgrading from windows 7 to windows 10 [21:14] It worked when I originally had set it up. [21:14] disabling smb1 cause it's so insecure [21:15] I have always used windows 10 [21:15] odd [21:15] maybe you didn't have smb1 disabled and now you do [21:16] there are a lot of settings I have to change in smb.conf to make it work for me [21:16] enable signing, enable spnego, disable the low quality password hashing [21:16] https://pastebin.com/BErX8fC9 [21:17] theres the testparm -s [21:17] smb encrypt=auto [21:17] server min protocol = SMB2 [21:17] client min protocol = SMB2 [21:17] client ntlmv2 auth = yes [21:17] ntlm auth = no [21:17] server signing = mandatory [21:17] client signing = mandatory [21:17] those are my basic settings [21:18] the smb.conf was just the default that was created and I just added what I needed at the end [21:19] yes, and that will be the default from that version of samba, that is pretty old, and it's defaults won't match what is in use today [21:19] all it takes is a single microsoft patch, and it's no longer compatable [21:19] that does not suprise me [21:19] do you clear out all those settings [21:24] DexDeadly: I don't have time to look into yours but mine works well with Windows 7 clients (older are not supported by design) https://paste.ubuntu.com/p/kjj72jb93w/ [21:26] thanks I'll take a look [21:26] I just added those lines you mentioned to the bottom of my smb.conf file [21:27] how does that work? [21:27] still no go [21:27] they need to be in the general section [21:28] maybe you need to start with the samba manual [21:28] ahh wait thats my issue [21:28] yea im gonna have to read the new way [21:28] i thought it was something simple but this may need redone [21:28] the signing = mandatory might cause you issues, but shouldn't [21:29] I require signing on all mine machines so no one can inject data [21:29] signing = mandatory is the new default (which is why it doesn't show up in testparm -s) [21:30] DexDeadly: are you sure samba wasn't upgraded without your knowing? zgrep samba /var/log/apt/history* [21:30] IIRC, the signing default value was changed due to the set of SMB vuln that were addressed not too long ago [21:31] sdeziel, shows up in mine [21:31] but I am on trusty on this server I'm looking at [21:33] patdk-lap: trusty has the same version xenial has [21:45] patdk-lap: I'm sorry, I was wrong on the "server signing" default value. [22:05] sorry i'm at work to lol let me get to this [22:17] nope took out server signing and still same thing === berglh_ is now known as berglh