[03:13] ahasenack: fyi, the catch-up script may not start until tomorrow (i'm going afk for the evening) [07:03] Good morning [07:10] hi lordievader, how are you? [07:10] Hey cpaelzer Doing good here, how are you? [07:11] same [08:36] weirdness [08:36] mars 14 09:32:43 hrmng systemd-networkd[963]: ipmi: Could not get LINKINFO: No data available [08:36] mars 14 09:32:43 hrmng systemd-networkd[963]: Could not set ifindex on netdev, ignoring: No data available [08:36] fails every boot. but systemctl restart systemd-networkd fixes it. [08:37] have other interfaces using the same driver that works fine. [08:38] No results found for "Could not get LINKINFO: No data available". [08:38] thanks google :p [09:12] uh, I moved it to another port, still failed. [09:12] gave it another name.. then it worked [09:14] odd. === Aztec03 is now known as SmokinGrunts === SmokinGrunts is now known as Aztec03 [11:00] rbasak: getting to the memcached review only now - with a potential lunch interrupt it might still take a while [11:00] are you super blocked onthat? [11:01] cpaelzer: no not blocked at all. [11:02] It can land "whenever" IMHO. No FFe required (again IMHO). [11:02] Just needs to make Bionic. [11:02] ok, so not postponing lunch then [11:02] Though of course earlier is better for wider testing as always. Yeah, of course don't postpone lunch :) [11:07] rbasak: do you have a ppa with that built already? [11:07] cpaelzer: yes. ppa:racb/experimental [11:23] rbasak: tests after lunch, the review on code is done and mostly ok [11:24] Thanks! [12:06] rbasak: hi, morning [12:06] rbasak: regarding my dfsg tarball question from yesterday [12:06] I saw your reply and nacc's [12:07] I don't know how to proceed now, if I should generate the dfsg tarball and go ahead of debian, risking its md5 being different from what debian will generate when the time comes, or doing nothing [12:08] mdeslaur: ^ this is a known thing that can happen, but when we have no choice we just do it, right? [12:09] * rbasak isn't sure if mdeslaur will be in today [12:09] that's a tough one [12:09] Oh. Hello :) [12:09] you can name it something else than dfsg [12:09] hi :) [12:09] is there precedence for that? [12:10] all of this because of some rfc files :/ [12:10] I believe I've done it before [12:10] Even if we name it something else, Launchpad will only accept that one orig for a given upstream version in Ubuntu, right? [12:11] I don't think so, all launchpad cares about is that the orig tarball doesn't have the same name as another [12:11] Oh, OK. [12:12] which package is this? [12:13] samba :-) [12:14] debian is at 4.7.4, no idea when they will update [12:14] we would like to go to 4.7.6 because of a corruption bug (fixed in 4.7.5) and the security fixes in 4.7.6 [12:14] yeah [12:15] changelog shows only bugfixes are in these new releases [12:15] so call it 4.7.6+adfsg [12:15] the secfixes have patches, but the corruption bug's patch is huge, about 100kb [12:15] and Andrew Bartlett himself is asking us to please use 4.7.6 [12:15] just make sure there's no "dfsg" versioning logic in debian/rules, I've seen that before [12:16] how will dpkg-buildpackage know which tarball to use? Where is that name referenced? [12:16] it finds the orig tarball name based on the version string in debian/changelog [12:16] ah, sure [12:16] Oh, of course. [12:17] will 4.7.6+dfsg be higher than 4.7.6+adfsg? [12:17] That's why what I said above is moot. The upstream version would be "different". [12:17] You could use 4.7.6+dfsg~ubuntu or something. [12:17] Then when Debian has one, we can "upgrade" the upstream version to it. [12:18] but you'll get a tarball collision if you do that [12:18] oh [12:18] ah yes, I think +dfsg~ubuntu-1ubuntu1 or similar would work [12:18] so [12:19] -0ubuntu1 please, to be consistent. [12:19] sorry, yes, 0ubuntu1 [12:19] samba-4.7.6+dfsg~ubuntu-0ubuntu1 [12:19] twice ubuntu? [12:19] Yes [12:20] But without the samba- prefix in the version string itself of course, but I think that's what you mean. [12:20] It's interesting that this isn't an already established pattern for us. [12:20] I wonder if there's a reason for that. It seems so obvious now. [12:20] do you guys know Mathieu Parent ? I would like to ping him to see if he is planning on an upload soon [12:20] so your tarball would be called samba_4.7.6+dfsg~ubuntu.orig.tar.gz [12:20] mdeslaur: right [12:21] And solves the problem pretty much entirely, in the case that we know that we're going ahead of Debian in the orig tarball construction. [12:21] I guess it doesn't happen often enough that anyone has documented the right approach [12:40] cpaelzer: memcached> thank you for the review! [12:44] yw rbasak [14:21] Ubuntu 16.04 Server, samba and kerberos working, smbnetfs is setup, ~/.smb/smb.conf is a copy of /etc/samba/etc/smb.conf customised for my domain/ realm but when smbnetfs is used, it only shows resources shared under WORKGROUP, not my current domain. [14:21] I also get this: [2018/03/14 12:58:22.380986, 0] ../source3/winbindd/winbindd_group.c:45(fill_grent) [14:21] My winbind version is 4.3.11, so it should not suffer from the bug mentioned in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=858601 [14:22] Debian bug 858601 in samba-vfs-modules "winbind: user authentication using windows domain fails after upgrade to 4.2.14+dfsg-0+deb8u4" [Important,Fixed] [14:24] I never used smbnetfs [14:26] ahasenack, smbnetfs is a kind of glue between fuse and mount.cifs [14:27] ahasenack, this should simplify mounting windows shares as a normal user. [14:27] and /etc/samba/smb.conf uses WORKGROUP, whereas your config in ~ uses another? [14:29] ahasenack, neither configfile mentions WORKGROUP, but smbnetfs seems to default to WORKGROUP. Ah, I see that my samba error message is not complete, my apologies: [14:29] [2018/03/14 14:01:26.482395, 0] ../source3/winbindd/winbindd_group.c:45(fill_grent) [14:29] Failed to find domain ''. Check connection to trusted domains! [14:30] ahasenack ^ [14:31] doesn't ring a bell, but winbind is notorious for errors here and there [14:32] ahasenack, the groups command returns the correct groups that this ADS user is a member of, which should also work through winbind, or is this assumption wrong here? [14:33] it's not wrong [14:33] I would try to verify this via other tools [14:33] ahasenack, do you have a suggestion which tools to use for testing this? [14:34] also modify the main smb.conf and use that one as the only source of configuration, to remove parsing multiple configuration files as a possible source to the problem [14:34] jlnl: smbclient, nautilus' network environment (forget what it is calleD) [14:35] smbclient eh, ok, I´ll read the manpages for that command. [14:43] ahasenack, smbclient works normally [14:45] jlnl: what is the version of the share server itself? [14:46] TJ-, Windows Storage Server 2012 Standard 6.2 [14:47] I wonder if the debug option used in this similar report might help, you may have the same issue https://sourceforge.net/p/smbnetfs/bugs/25/ [14:51] TJ-, that might very well be the case. I couldn´t find any similar solution that automatically iterates through available shares though. [14:53] TJ-, Look at the year: 2011, that´s ancient... [14:53] Still open. [14:53] jlnl: or it was not confirmed as an issue, user solved it and never reported back on how :) [14:54] TJ-, yes, that unfortunately happens ... a lot :-/ [14:56] TJ-, judging by the last entry, this shpac did not have any password authentication setup on the server (s)he was connecting to. Our network is a bit more secure than that :-) [15:08] rbasak: deb maintainer told me they have a 4.7.6 tarball in their "pristine-tar" branch, any idea where that would be? I checked salsa and it's still at 4.7.4 in the pristine-tar branch [15:14] ahasenack: perhaps the maintainer forgot to push it? [15:14] Or it got pushed to an old alioth repo? [15:14] If you don't see a branch called "pristine-tar" that has been updated recently, it's not been pushed there. [15:15] mdeslaur: are you planning on doing the php7.0 update still? i've gotten another offline ping re: security update (this time for 7.0.28) [15:16] nacc: hi! 7.0.25 only fixed "low" cves, so I didn't release it as a security update [15:16] nacc: are you planning 7.0.28 updates? let me look at the cve list, one sec [15:17] rbasak: yeah, maybe alioth, will check there too before responding [15:17] nacc: looks like 7.0.28 has "medium" CVEs, so that one we should release as a security update [15:18] mdeslaur: err, it was 7.0.27 that was waiting for you, iirc [15:18] mdeslaur: i can prep 7.0.28 for you today [15:18] oh, you had 7.0.27? [15:18] (or 7.0.26, i can't remember) [15:18] darn, sorry, did I drop the ball on that one? [15:18] yeah, it was some not yet published version [15:18] mdeslaur: it's alright :) [15:18] I thought the one you wanted me to look at was the one that went to -updates [15:19] mdeslaur: sorry, miscommunication on my part [15:19] nacc: so we need 7.0.28 and 7.1.15 [15:20] and 7.2.3 [15:21] mdeslaur: ack [15:22] nacc: if you don't mind preparing 7.0.28 and 7.1.15, ping me and I'll push them straight to -security [15:22] nacc: I updated LP: #1744148 but it only covers the 7.0 branch. Would it help if I add the 7.1 and 7.2 branches? [15:22] Launchpad bug 1744148 in php7.0 (Ubuntu Xenial) "[MRE] Please update to latest upstream release 7.0.28" [Wishlist,In progress] https://launchpad.net/bugs/1744148 [15:23] mdeslaur: yep, will do [15:23] sdeziel: if the CVEs are the same, probably [15:23] nacc: OK I'll check if there is an overlap [15:24] sdeziel: thanks -- otherwise, it probably makes sense to have distinct bugs [15:25] mdeslaur: and yeah, i plan on doing the 7.2 update today anyways (i believe it might fix a bug someone is hitting) [15:35] nacc: they all overlapped so I extended the existing bug [15:38] sdeziel: thank you! [15:39] that's the least I can do, I'm pretty glad that you take such good care of PHP :) [16:59] TJ-, thank you for the useful pointers. I´ve got to go now. [17:11] with SELinux, if you are running as a low priviledged user and get an "invalid argument" error when you try to chcon a file, is this due to permissions? [17:12] I am following the selinux wiki example [17:12] https://selinuxproject.org/page/Guide/Contexts > to what I refer [17:13] sudormrf: are you using selinux on ubuntu? [17:13] I am testing it [17:14] sudormrf: given that the selinux package in ubuntu hasn't been updated since 12.04, i'm not sure how well it is supposed to work [17:14] ok, that's fine. curious about the "error" I am seeing [17:14] I am _betting_ it's permissions [17:14] sudormrf: might be easiest to ask a selinux channel? [17:15] oh, didn't know there was one :D [17:15] sudormrf: or mailing list, or whatever [17:16] thanks :) [17:44] ahasenack: running the samba import now [17:53] Google failed me. My search was "why was php7.1 not included in bionic" [17:54] i'm sure theres a better reason than "we have 7.1 now" [17:54] UGH 7.2 i mean [17:56] MitchT: we don't want to have two versions of php in the archive [17:56] makes sense [17:57] kinda rough for all those magento users out there scrambling for an mcrypt fix, but honestly we should have been using openssl by now. I know there was a php7.1x bug, but still [17:57] thank the maker for pecl. [17:58] MitchT: afaik, magento is not in ubuntu [17:59] oh absolutely, no more likely than wordpress to be in it [17:59] its an ecommerce platform [17:59] but its still using mcrypt, and that lib's deprecation and omission has left a lot of us coming up with "creative fixes" [17:59] ;) [18:00] also, being an encryption library, you might understand how its scary to just swap it out for something else [18:01] MitchT: it's equally scary to be using something deprecated :) [18:02] * MitchT agrees [18:03] MitchT: seems like it would be equally easy to run magento on php7.0 on 16.04 in a VM or container [18:04] it would be.. and that was an option, but i really want to go with 18. our current site was completed in 2011 [18:04] granted there have been a lot of adjustments, but its still running ubuntu 14 [18:04] our IT dept is small and we have a lot of projects [18:05] MitchT: you can still run 18.04 as the host [18:05] i am [18:05] i use php7.2 and pecl install mcrypt [18:05] makes magento very happy on a tiny azure vm [18:06] hehe... hope Odd_bloke / team gets that azure image bug fixed soon. My automated deployments are locked to a build from Feb. [18:06] ubuntu 18 has exceeded my expectations by a longshot, i've not used something that works as well as this in a LONG time. === MartesZibellina is now known as FalconMillennium [20:09] ok, can someone explain how two dhclient processes would be running for a single interface, resulting in flood to the dhcp server... ? [20:09] Onepamopa: cloud workload? [20:10] what the .... is that? [20:14] I'll take that as a no [21:13] MitchT: yeah that works too -- tbh, based upon xevious' recommendation and thinking about it, i think long-term, we want to drop everything but the interpreter, pecl and composer from ubuntu at some point -- it's just not possible to keep everything current [21:27] rbasak: ahasenack: fyi, the samba patches-applied import fails due to https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=850843 [21:27] Debian bug 850843 in dpkg-dev "dpkg-source in stretch cannot extract samba_3.6.5-2.dsc" [Important,Fixed] [21:27] which is parallel true for any modern dpkg-source in ubutnu relative to the one that was in precise (apparently) [21:42] ahasenack: i believe samba patches-unapplied has been reimported fully now [21:53] rbasak: for the git-ubuntu merge fix (LP: #1734364), I think I'll need the tags support, as well, in order to create a Repo (was CommitGraph) object to pass in, write it to the repository, and then do a lookup of some objects in that repository (I don't believe the placeholder exists once we've written it, right?) [21:53] Launchpad bug 1734364 in usd-importer "merge start fails with bind9" [Critical,In progress] https://launchpad.net/bugs/1734364 [22:10] rbasak: http://paste.ubuntu.com/p/VSbYQqTPhp/ for reference to what i mean, hjopefully [22:54] sorry, jst asked this on ubuntu but I beleive it fits here more. I'm trying to run a python server on ubuntu using systemd to be reverse-proxied by nginx. I'd like to automatically restart the server every day or week or some arbitrary amount of time... if I use WatchdogSec= in [Service] of the unitfile, will this just kill the process or send a SIGINT? [22:55] i have to think you'd be better served by a cronjob entry [22:55] jayjo, ya, you are likely to get better replies here, though I'm not sure of the right unit incantation [22:56] same, it would be trivial there [22:56] so a service to start it and a cronjob to restart it? [22:57] I mean, *best* would be to sort out whatever resource consumption problem you've got that makes a restart seem like a good idea :) but a cronjob to run something like systemctl restart foo.service can't be so bad [23:17] rbasak: ah i think we want to update our pytest config to be xfail_strict=true (which will make xpass also a fail) [23:58] need some help with this.. logger -p mail.info writes to /var/log/mail.log fine, but not sending to remote IP .. mail.info @192.168.1.100 is there [23:59] 14.04 [23:59] axisys: does *anything* make it to that remote syslog server?