[06:31] <cpaelzer> good morning
[09:45] <TariusWork> good morning : This is highly relative because it depends on user time zone :D :D ... And I think that this is international channel for many peoples from many time zones.
[09:50] <albech> Anyone here running Bareos backup? I am a little worried about that they only release packages with the initial version and all updates are through Git unless you pay for the commercial version. Wondering how many issues people are experiencing with the community version.
[10:03] <andol> albech: Well, looks like the GitHub repository contains their ./debian packaging folder, so ought to be fairly trivial to build your own packages whenever there is a new minor release. Then I guess there is also the Bareos packages in the regular Ubuntu archives, but I have no experience with how well they are kept up to date.
[10:16] <albech> andol: cheers.. Guess id have to have a buildbox and move compiled package to production. Looked a little at the pricing and there is very little change that i can convince management to go from our free self written backup solution to a paid solution in the price range they are asking :(
[11:21] <ren0v0> Hi, can anyone help me as to why i'm getting this log in cron service?   Dec 18 10:17:01 prod-lb CRON[20762]: pam_unix(cron:account): expired password for user root (root enforced)
[11:23] <ahasenack> ren0v0: I suppose some cron script is using su or sudo to switch to root, and that is triggering that that log
[11:23] <ahasenack> you should be able to see which script in the same log
[11:30] <ren0v0> ahasenack, no there is nothing about the script, but the script works fine manually run and in cron.daily
[11:30] <ahasenack> did you check /var/log/syslog or /var/log/auth?
[11:30] <ren0v0> the issue here is as you can see from this paste > https://pastebin.com/tjFtwj1f  > that ubunt has different line entries for daily and hourly. hourly doesn't use "anacron", is this a mistake?
[11:35] <ahasenack> ren0v0: I think the anacron is fine
[11:35] <ren0v0> ahasenack, my point is anacron isn't used in the cron.hourly job, why?
[11:36] <ahasenack> anacron is meant to "catch up" cron jobs that didn't run when the system was off
[11:36] <ahasenack> an hourly cron job that didn't run is ok to run in the next hour, so no need for anacron to "speed it up"
[11:36] <ahasenack> that's my thinking
[11:38] <ren0v0> ahasenack, hmm interesting
[11:38] <ren0v0> but, could the lack of this anacron pipe cause this auth issue ?
[11:38] <ahasenack> do you have a password for root?
[11:39] <ahasenack> it's complaining the password is expired, maybe all you need to do is reset it, if that's part of your password policy
[11:39] <ren0v0> ahasenack, sure i can set one, but ubuntu doesn't have one by default does it ?
[11:39] <ren0v0> i'm sure it does on ubuntu server though
[11:39] <ren0v0> ahasenack, the weird thing is that cron.daily runs fine, which is why it confused me
[11:40] <ahasenack> you have the same job in cron.hourly and cron.daily?
[11:43] <ren0v0> nope
[11:44] <ren0v0> i moved it
[11:45] <ahasenack> try pasting a bit more context from syslog around that expired password line
[11:46] <ren0v0> ahasenack, you might be right about the password expiring
[11:46] <ren0v0> i usually change it, but maybe this time i didn't, and this is a droplet so i think theyn might expire root passwords
[11:46] <ahasenack> well, it's what the log is saying :)
[11:47] <ackk> rbasak, hi, do you see any reason ubuntu-advantage-tools couldnt' use bash rather than dash?
[12:07] <ren0v0> Hi, i thought that open files were limited by the figure in "ulimit -n", for me that is "163840". however, if i run "lsof | wc -l", the figure is greater that this number, why is that ?
[12:09] <ahasenack> ren0v0: ulimit is not global
[12:09] <ahasenack> ren0v0: the global setting is somewhere in /proc
[12:10] <ren0v0> cat /proc/sys/fs/file-nr    ?
[12:11] <ahasenack> I don't remember from memory
[12:11] <ren0v0> well this shows far less at 7000, not 180,000
[12:11] <ren0v0> so not sure what the difference is
[12:13] <ahasenack> I think file-max is a better choice
[12:13] <ren0v0> this is like 2million
[12:58] <ackk> ren0v0, file-nr is a read-only attr, it shows the current value (see man proc)
[13:02] <ren0v0> ackk, that's useful also, as its far different than "lsof | wc -l"
[13:02] <ren0v0> like, 150,000 difference
[13:31] <rbasak> ackk: as long as the package depends on it, I believe it's OK.
[13:31] <rbasak> ackk: though you might be the only thing depending on it in a minimal environment, so people may not like that. Which seed are you in again?
[13:40] <ackk> rbasak, ubuntu-minimal
[13:40] <ahasenack> cpaelzer: these are the changes I noticed when rebuilding libzstd with dh9 compared to dh10: http://pastebin.ubuntu.com/26208547/ (on xenial)
[13:40] <ackk> rbasak, u-a-tools don't depend on dash fwiw
[13:40] <ackk> as it's Essential: yes
[13:40] <ahasenack> cpaelzer: seems ok
[13:41] <ackk> rbasak, it doen't have any Depends: ATM
[13:41] <rbasak> ackk: is it possible to install only ubuntu-minimal without bash installed?
[13:41] <rbasak> If it is, then by switching to bash you'd be imposing that upon ubuntu-minimal.
[13:43] <cpaelzer> ahasenack: reading the link
[13:44] <cpaelzer> ahasenack: yep LGTM as well
[13:44] <ahasenack> thx
[13:45] <ahasenack> vimdiff is nice for this kind of thing, thanks for the suggestion
[13:45] <cpaelzer> yw
[13:52] <ahasenack> rbasak: dmb meeting is in ~1h8min?
[13:54] <rbasak> ahasenack: yes. Thank you for the reminder!
[13:56] <ahasenack> cheers
[13:56]  * ahasenack -> lunch
[14:06] <ackk> rbasak, isn't it the same as dash?
[14:07] <ackk> rbasak, since both dash and bash are Essential, they have to be installed anyway, right?
[14:08] <rbasak> I didn't think bash was Essential?
[14:08] <rbasak> If it is, then sorry, you're right.
[14:08] <ackk> yeah I'm slightly surprised it is
[14:08] <ackk> but it is
[14:09] <rbasak> Then no reason for a dependency as I think you said, and sure, use it.
[14:09] <ackk> rbasak, cool, thanks for confirming
[14:09] <rbasak> The only other reason I know about is performance (eg. startup scripts), but I don't think that's an issue for this tool.
[14:09] <ackk> right, it's not something that's even run automatically
[14:47] <ahasenack> ackk: is bash Essential since precise?
[15:10] <dpb1> right, that was my next question
[15:10] <dpb1> :)
[15:51] <boxrick> Is there anyway to suppress the Ubuntu preseed whine about no gateway
[15:51] <boxrick> My network has no gateway, so this warning needs to just accept that and carry on.
[15:52] <ahasenack> it's unusual, so it sounds like a good warning to have
[15:53] <boxrick> Well, this is a fully automated install. So any popup is complete disaster.
[15:53] <ahasenack> ah, it stops the install?
[15:53] <boxrick> Yea
[15:53] <ahasenack> ok, didn't know that
[15:53] <ahasenack> can you set the gateway to your own ip?
[15:53] <rbasak> I think you can bypass it, but I don't recall the detail right now.
[15:53] <boxrick> I can set it to anything but it seems a bit of a sad fix
[15:54] <ahasenack> maybe set it just to satisfy the installer and later remove it via another step in your automation?
[15:54] <rbasak> The warning is just another debconf prompt whose answer you can preseed.
[15:54] <rbasak> I think.
[15:54] <ahasenack> just thinking about workarounds, sorry
[15:54] <boxrick> No its good thanks :)
[15:54] <rbasak> You just need to figure out the name of the prompt I think.
[15:54] <boxrick> Why apologise for helping! Any ideas how I can find the name of the prompt?
[15:55] <rbasak> It'll be in a debconf template somewhere
[15:55] <rbasak> Somewhere in the source.
[15:55] <rbasak> I can't remmber the package name
[15:55] <rbasak> The package that does network configuration in the installer
[15:55] <rbasak> netcfg maybe?
[15:56] <rbasak> Here we are
[15:56] <rbasak> debian/netcfg-dhcp.templates in the netcfg package
[15:56] <rbasak> http://paste.ubuntu.com/26209463/
[15:56] <rbasak> netcfg/no_default_route
[15:56] <rbasak> Type: boolean
[15:57] <rbasak> Try setting that to yes or true or similar.
[15:57] <rbasak> Assuming the message against that key in the template file is the message you see?
[15:58] <TJ-> wouldn't it be netcfg/get_gateway=none  ?
[15:58] <rbasak> Maybe. I'm just guessing!
[15:58] <TJ-> me too!
[16:00] <rbasak> I suspect get_gateway might make it ignore any DHCP response. And no_default_route will make it proceed without a gateway. But again, I'm just guessing :)
[16:00] <rbasak> ignore any DHCP response> I mean ignore the gateway part only, of course
[16:05] <boxrick> Yea will have to try
[16:05] <boxrick> From my experience, any static config is complete
[16:05] <boxrick> Not just small sub components
[17:23] <tomreyn> is there an EOL for 12.04 ESM? if so, which? also, will there be ESM for 14.04?
[17:23] <tomreyn> sorry, i meant to ask this in -hardened
[22:15] <jiffe> so upgrading ubuntu seems to have broken python (and everything in the upgrade that depended on python)
[22:17] <tomreyn> ok
[22:20] <jiffe> I'm not sure how to fix this, asking in #python they just say its an ubuntu problem
[22:21] <jiffe> it fails at Exception: python3.5 -c 'import imp; print(imp.get_tag())' failed with status code 1 with ImportError: No module named '_sysconfigdata_m'
[22:21] <jiffe> but that module does seem to exist
[22:22] <tomreyn> describe what you were doing, show input and output, explain which ubuntu release you are working on and which major modifications you have made.
[22:23] <tomreyn> !paste | jiffe
[22:24] <jiffe> I had to add /usr/lib/python3.5/plat-x86_64-linux-gnu/ to the python path and then apt-get upgrade finished
[22:24] <jiffe> I don't know if that path is missing from something
[22:25] <dpb1> jiffe: what ubuntu release, and what were you upgrading?
[22:25] <dpb1> jiffe: often this information can be pastebined from /var/log/apt/...
[22:27] <jiffe> just doing an apt-get upgrade on Ubuntu 16.04.3 LTS
[22:27] <dpb1> and what package failed to upgrade.
[22:27] <jiffe> these are the packages it was installing: apport bind9-host distro-info-data dnsmasq-base dnsutils dpkg dpkg-dev grub-legacy-ec2 klibc-utils libbind9-140 libdns-export162 libdns162 libdpkg-perl libisc-export160 libisc160 libisccc140 libisccfg140 libklibc liblwres141 libpulse0 lxcfs lxd lxd-client mdadm python3-apport python3-problem-report ubuntu-minimal ubuntu-server ubuntu-standard
[22:27] <jiffe> first error occurred after python3-problem-report
[22:27] <dpb1> ok
[22:30] <dpb1> jiffe: dpkg -l |grep python-problem-report?
[22:32] <jiffe> interesting, its not in dpkg -l
[22:34] <jiffe> oh whoops, missing the 3, its there
[22:34] <jiffe> ii  python3-problem-report                                                                                            2.20.1-0ubuntu2.14                                            all                                                           Python 3 library to handle problem reports
[22:35] <dpb1> and I take it you applied the workaround already?
[22:35] <dpb1> therefor you are on the more updated version.
[22:36] <jiffe> yes everything installed after fixing the path
[22:37] <jiffe> the global path is still broke, I only fixed it for that shell's session
[22:37] <jiffe> opening a new ssh session python3.5 is still failing for the same reason
[22:42] <jiffe> also I don't know why I'd be the only one affected here, I don't know that people are using python on this machine
[22:42] <jiffe> certainly not python3.5 so that should all be default
[22:43] <dpb1> jiffe: what does yours say?
[22:43] <dpb1> http://paste.ubuntu.com/26211504/
[22:44] <sarnold> what's he sys.path on a fresh shell? echo 'import sys; print(sys.path)' | python3.5
[22:44] <sarnold> hah, I see dpb1 beat me to it :)
[22:44] <dpb1> jiffe: in the past when things have bitten me like this, it's usually that I have pip modules installed that are misbehaving, I"m using a custom version of python3 that I compiled it myself, etc.
[22:46] <jiffe> I can't even run python3.5 because of this path issue, even running python3.5 with no arguments fails, python3's path lists ['', '/usr/lib/python35.zip', '/usr/lib/python3.5', '/usr/lib/python3.5/plat-x86_64-linux-gnu', '/usr/lib/python3.5/lib-dynload', '/usr/local/lib/python3.5/dist-packages', '/usr/lib/python3/dist-packages']
[22:46] <dpb1> `which python3`
[22:47] <jiffe> python3 is /usr/bin/python3 and python3.5 is /usr/local/bin/python3.5
[22:47] <dpb1> well
[22:48] <dpb1> /usr/local/bin/python3.5 smells of a compiled version of python on the system indeed
[22:48] <dpb1> (user-compiled, I mean)
[22:48] <dpb1> ls -ld /usr/bin/python3  ?
[22:50] <jiffe> /usr/bin/python3 -> python3.5
[22:50] <dpb1> ls -ld /usr/bin/python3.5
[22:50] <jiffe> /usr/bin/python3.5
[22:50] <jiffe> binary no symlink
[22:51] <dpb1> ok
[22:51] <jiffe> not sure where /usr/local/bin/python3.5 came from, nothing compiled on this machine
[22:51] <jiffe> binary timestamp is july 2016 and this machine was built this year
[22:51] <jiffe> couple months ago
[22:52] <dpb1> what's your sha1sum
[22:52] <dpb1> on my two xenial systems it's 3eab1d3d6cafd34315183305819cbbc361ded361  (/usr/bin/python3.5)
[22:52] <dpb1> (amd64)
[22:53] <jiffe> that matches /usr/bin/python3.5, different sum for /usr/local
[22:53] <sarnold> try dpkg -S /usr/local/bin/python3.5. Maybe you've got a gross package installed?
[22:55] <jiffe> yeah I don't know what this is or why its here but that's the problem, I renamed it and python3.5 works fine from /usr/bin
[22:56] <dpb1> ok
[22:56] <dpb1> ya, I've had that before