[01:05] <an3k> why is PATH="$HOME/bin:$PATH" fine but LD_LIBRARY_PATH="$HOME/lib" is not?
[01:39] <joeb3_> LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$HOME/lib"
=== quantic_ is now known as quantic
[01:54] <sarnold> an3k: because you're potentially changing the behaviour of -all- programs you run if you're defining a new location to find symbols; but you're unlikely to have funny versions of existing program names in ~/bin/
[01:54] <sarnold> an3k: you're right that it's not -that- different, but it does potentially affect nearly every program you'll run.
[01:55] <an3k> hmm, when I login and do echo $LD_LIBRARY_PATH it's empty, same for the other env path i set
[01:56] <an3k> and doing "LD_LIBRARY_PATH="$HOME/lib"" in .profile doesn't work but export LD_LIBRARY_PATH="$HOME/lib" does
[01:57] <an3k> I know that export is only for the current login and not permanently (as long as you don't put it in the .profile :) but afaik PATH="/foo" is temporary too?!
[01:59] <tarpman> anything you set in your shell applies only to that shell process
[01:59] <tarpman> export means it gets copied to child processes as well
[01:59] <sarnold> an3k: something else has already done an 'export PATH' somewhere during login; I never did figure out the full 'export' rules in shells; check the output from export -p | grep -e PATH -e LD_   to se...
[01:59] <tarpman> PATH=... applies to your current shell only; export PATH=... applies to your current shell _and_ any processes it spawns
[02:00] <tarpman> hmm, PATH might be a bad example for that.
[02:01] <an3k> so with PATH I can do "binary" and it checks $HOME/bin for that but if "binary" wants to start "foobarnary" which is also in $HOME/bin then it doesn
[02:01] <an3k> 't work?
[02:02] <sarnold> since PATH is exported, whatever you start will get the value..
[02:02] <an3k> and if I want that to work I would have to export PATH
[02:02] <an3k> ah yeah, that's why the bad example :)
[02:02] <an3k> but I got the idea. Thank both of you!
[02:03] <tarpman> but if you do FOO=bar in your current shell, spawned processes won't see it unless you explicitly export it
[02:03] <tarpman> PATH is a bad example because it's already exported, as sarnold said
[02:03] <sarnold> tarpman: I thought I understood all this but then:
[02:03] <sarnold> $ FOO=bar /usr/bin/env | grep FOO
[02:03] <sarnold> FOO=bar
[02:04] <tarpman> sarnold: difference between that and $ FOO=bar; /usr/bin/env | grep FOO
[02:04] <sarnold> tarpman: dude. 22 years I've been doing this..
[02:04] <sarnold> tarpman: thanks :D
[02:05] <tarpman> :D
[02:05] <an3k> very nice we've talked about :D
[02:06] <sarnold> an3k: definitely. I've wondered just what little bit I've missed for a decade or something but never really investigated. tarpman points it out with a one-char change to my example.
[02:09] <an3k> hehe yeah. This is also difficult to google for and I often noticed that someone has to ask the right question at the right time in right location :)
[02:13] <sarnold> that can be quite difficult too :)
[02:29] <an3k> grrrr su vapoursynth -c "echo $HOME" = /root
[02:58] <tarpman> an3k: the difference between su -c "echo $HOME" and su -c 'echo $HOME'
[03:00] <an3k> oh my ... why, just why?! :)
[03:00] <sarnold> tarpman: lol, two chars this time :)
[03:00] <tarpman> ^_______________^
[03:00] <sarnold> tarpman: and, uh, I looked right at it and missed it. sigh. it's a good thing this is a three-day weekend. :)
[03:01] <tarpman> sarnold: we were talking about that in scrum this morning, actually. turns out my cow-orkers get to work on monday
[03:01] <tarpman> (i'm in canada, so don't get a stat until next month)
[03:02] <tarpman> sarnold: basically: enjoy your long weekend :)
[03:02] <sarnold> tarpman: hmm, sorry about the CAD, I'm doing my best to burn fuel though I swear :)
[03:02] <tarpman> ^^
[03:02] <sarnold> though I realize i could be doing more to consume more maple syrup.
[03:02] <tarpman> an3k: for the record, a bunch of scripting languages inherited the ' vs " thing from sh, so it's good to be alert about
[03:05] <an3k> yeah, I know that from PHP but there it has a complete different meaning (int/var) but the behaviour in bash is very confusing.
[03:07] <tarpman> in terms of performing substitution or not, I meant
[03:07] <tarpman> not sure what you mean by int/var?
[03:08] <an3k> in PHP bla='5' is correct (integer) as well as foo="more" (variable) and vice-versa it's wrong
[03:08] <an3k> What I can't understand is that su user -c "echo $HOME" is executed as root and su user -c 'echo $HOME' as user.
[03:09] <sarnold> the "" string is interpolated by the shell
[03:09] <sarnold> the '' string is not
[03:09] <tarpman> "echo $HOME" - the $HOME part is evaluated in your current shell and the result passed to su
[03:09] <sarnold> it isn't about root execution or user execution
[03:09] <tarpman> su user -c 'echo $HOME' the entire string is passed to su as is, and evaluated by the shell invoked by -c
[03:11] <an3k> that is a very informative evening :) Thank you
[03:12] <tarpman> an3k: were you the person I was talking about l-smash and static linking with? I still have a schroot around here somewhere with it half installed ...
[03:12] <an3k> yes, must be me :)
[03:15] <an3k> do you need help with?
[03:18] <an3k> hmm, obviously ~/ is always evaluated by the shell invoked by -c
[03:18] <sarnold> depends upon '' vs "" again
[03:20] <an3k> I have su vapoursynth -c "~/vs_user.sh" and added echo in front of it. I always got ~/vs_user.sh echo'ed with " as well as '
[03:20] <tarpman> no, he's right - " counts as quoting in the case of ~
[03:20] <tarpman> sh(1) says only $ ` " \n are evaluated inside "
[03:21] <tarpman> pardon, $ ` \
[03:21] <sarnold> w.t.f. 22 years..
[03:21]  * sarnold hides in shame.
[03:21] <sarnold> tarpman: I hope you're keeping track of how many beers I owe you.
[03:21] <tarpman> heh. I have the manpage in front of me, it's not like I have it memorized :P
[03:21] <sarnold> surely were into the dozens by now :)
[03:22]  * tarpman writes this down
[03:22] <tarpman> sarnold: planning to attend debconf!? :D
[03:22] <tarpman> er, the next one in north america, anyway...
[03:22] <sarnold> tarpman: heh, one of these years..
[03:22] <sarnold> I wanted to make it to last year's, but we couldn't convince management that Yet Another person needed to go.. :(
[03:23] <an3k> thank god ~ isn't evaluated otherwise my build_script may be broken :)
[03:27] <sarnold> have a good weekend tarpman, an3k :)
[03:27] <tarpman> ttyl :)
[03:27] <an3k> bye sarnold, you too!
[03:28] <tarpman> an3k: right. so I think I'm at the same point you were ....  /usr/bin/ld: /usr/local/lib/liblsmash.a(description.o): relocation R_X86_64_32 against `.bss' can not be used when making a shared object; recompile with -fPIC
[03:29] <an3k> As far as I can remember I solved it by compiling l-smash with -fPIC (CFLAGS)
[03:31] <an3k> PATH="$HOME/bin:$PATH" PKG_CONFIG_PATH="$HOME/lib/pkgconfig" ./configure --prefix="$HOME" --extra-cflags="-I$HOME/include -fPIC" --extra-ldflags="-L$HOME/lib"
[03:32] <tarpman> aha, there it is
[03:32] <tarpman> for l-smash: ./configure --disable-static
[03:33] <tarpman> then everything works.
[03:33] <tarpman> modulo your various $HOME stuff - i'm in a chroot so installing everything to /usr/local is acceptable
[03:33] <tarpman> mystery solved :)
[03:33] <an3k> oh, you don't care about static links?! then I can give you my build script :)
[03:33] <tarpman> heh, I don't care about l-smash either ;) I was curious how you got to where you did (thinking you had to rebuild gcc)
[03:34] <tarpman> audio/video is not my area one bit
[03:34] <an3k> ah hehe
[03:34] <an3k> well, I even don't have to rebuild gcc
[03:34] <tarpman> didn't think so ;)
[03:34] <tarpman> glad you sorted that :)
[03:34] <an3k> a very knowledgable gcc dev told me how I can do it without
[03:34] <tarpman> but now you know, --disable-static rather than --extra-cflags
[03:35] <an3k> yeah. I built everything without static / with shared and it fully works
[03:35] <tarpman> awesome
[03:35] <tarpman> good place to be
[03:35] <an3k> if you just run this in the VapourSynth directory you should have a static lib for L-SMASH Works
[03:36] <an3k> ar cr libvslsmashsource.a lsmashsource.o video_output.o libavsmash_source.o lwlibav_source.o ../common/utils.o ../common/qsv.o ../common/libavsmash.o ../common/libavsmash_video.o ../common/lwlibav_dec.o ../common/lwlibav_video.o ../common/lwlibav_audio.o ../common/lwindex.o ../common/video_output.o
[03:36]  * tarpman kills off that particular schroot
[03:36] <an3k> maybe change "ar cr" to "ar csr"
[03:37] <an3k> that's what he said about how to avoid rebuilding gcc :)
[04:52] <LyF3L3SS> whats up all
[05:55] <NoCode> How do you upgrade to another release if the VPS is headless, and I'm connecting to it via ssh? Says it's not recommended to upgrade via ssh. But then it'll tell me that there will be another port to connect to if something goes array in which port 1022 will be used and thought I seen a mention of screen there.
[05:56] <NoCode> Should I use screen anyway? I guess it would make sense to use screen instead of a proper ssh terminal. That way I can reattach
[05:56]  * NoCode is thinking out loud.
[06:05] <NoCode> Should I keep /etc/rsyslog.conf when upgrading?
[06:19] <an3k> I'm awake but I have absolutely no clue. Personally I would keep rsyslog as long as I don't have to care about it's size / free diskspace
[09:30] <lordievader> Good morning.
[10:02] <arcsky> hi,  sudo vconfig add eth0 100 means the eth0.100 got vlan-id 100 ?
[13:38] <buengenio> good day
[13:38] <buengenio> installed the 4.4 (and tried the nightlies) and after reboot get "gave up waiting for root device"
[13:38] <buengenio> the grub menu entries look exactly the same as my working 4.3 kernel (i.e. root=UUID..... is all there..)
[13:38] <buengenio> any ideas?
[13:43] <lordievader> What are you trying to mount? What fs type?
[13:43] <buengenio> ext4
[13:44] <buengenio> just trying to boot into 4.4 kernel
[13:44] <lordievader> What version of Ubuntu and where did you get the 4.4 kernel from?
[13:44] <buengenio> installed official ubuntu mainline packages
[13:44] <buengenio> http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.4-wily/
[13:44] <lordievader> The ext fs is not on lvm or luks or something?
[13:45] <buengenio> no, it's not LVM
[13:46] <buengenio> basically, the stock 4.2 that came with wily - boots, 4.3 - which I also installed from the same ppa -boots, 4.4 (and 4.4-nightlies) - don't boot
[13:46] <buengenio> same root UUID
[13:46] <buengenio> can I try the good old root=/dev/nvme0n1p2 method?
[13:46] <lordievader> What method is that?
[13:47] <lordievader> Is there something in the 4.4 kernel that you really need by the way?
[13:47] <buengenio> yes - proper i915  suppoprt
[13:48] <buengenio> currently getting a lot of drm errors and timeouts during boot
[13:48] <buengenio> apparently it was solved in 4.4
[13:48] <buengenio>  *ERROR* dp aux hw did not signal timeout (has irq: 1)!
[13:48] <buengenio>   32.770232] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training
[13:48] <buengenio> [   33.538109] [drm:intel_dp_complete_link_train [i915]] *ERROR* failed to start channel equalization
[13:48] <lordievader> A video card in a server?
[13:49] <buengenio> yes
[13:49] <buengenio> why is that unusual?
[13:49] <buengenio> home server :)
[13:49] <lordievader> Most servers are headless.
[13:49] <buengenio> not this one
[13:50] <lordievader> buengenio: Do you get busybox when it gives up waiting for the root fs?
[13:50] <buengenio> i think so
[13:50] <buengenio> yes
[13:51] <lordievader> buengenio: Can you mount the root-fs manually in there?
[13:51] <buengenio> haven't tried
[13:51] <buengenio> but it also says /dev does not exist
[13:52] <buengenio> so I suppose I wouldn't be able to mount anything
[13:52] <lordievader> Have you checked it's existence?
[14:16] <NoCode> http://pastebin.com/E3AtcKaE -- Hi, would anyone know a solution? I'm not sure if such errors in the log are responsible for my index page being blank.
[14:35] <lordievader> Probably, a dependecy cannot be started. That is not the only error I see though...
=== Piper-Off is now known as Monthrect
=== unreal_ is now known as unreal
[17:13] <YamakasY> what does the dhcp failover mechanism failover ? the leasesfile ?
[17:26] <patdk-lap> no
[17:26] <patdk-lap> you must configure it in the config on each dhcp server
[17:42] <jonah> hi does anyone know what to insert into csf/iptables to allow samba shares on local network? Or can anyone please help so I can access samba, my firewall is blocking it. If i turn off the firewall it works great but I'd like to keep things secure and have the firewall running
[17:43] <jonah> I only need samba on my local network
[17:43] <jonah> any help would be great. thanks
[17:46] <OerHeks> jonah, maybe this page is a help >> http://askubuntu.com/questions/649969/what-ports-does-the-samba-client-require
[17:47] <jonah> OerHeks: hi thanks that's great. it was also the rules set I was looking for. I've not used iptable rules before so just wondered how I put the rules set in there and make sure it's right?
[17:47] <OerHeks> jups, run: netstat -tulpn | egrep "samba|smbd|nmbd|winbind"  # and these ports are your answer.
[17:48] <jonah> OerHeks: thanks
[17:50] <YamakasY> patdk-lap: ok, but it will never serve the same address from my backup dhcp it seems ?
[17:51] <jonah> OerHeks: it's this bit I don't really get, the syntax... http://www.perturb.org/content/iptables-rules.html
[17:51] <jonah> OerHeks: how do I make a few rules that allow my local network?
[17:55] <patdk-lap> YamakasY, that depends on how you configure it
[17:55] <jonah> I've posted here for advice: http://ubuntuforums.org/showthread.php?t=2310138&p=13424118#post13424118
[17:55] <jonah> thanks again for any help
[17:55] <patdk-lap> and if your client supports bootp or real dhcp
[17:55] <jonah> bye for now
[17:56] <YamakasY> patdk-lap: yeah it does but I have set a split on 255 on the master
[17:56] <YamakasY> and I assume it should takeover when I stop it
[18:09] <patdk-lap> isc dhcp doesn't takeover
[18:09] <patdk-lap> they will always work
[18:10] <patdk-lap> except the slave will be delayed like a second
[18:10] <patdk-lap> for bootp, you must configure a range for each to use
[18:10] <patdk-lap> for dhcp, they will split the pool between the two
[18:10] <patdk-lap> and update each other on usage
[18:21] <YamakasY> patdk-lap: ok, but what about dns than ? that is kinda odd
[18:23] <patdk-lap> heh?
[18:23] <patdk-lap> what does dns have to do with dhcp or failover?
[18:23] <YamakasY> patdk-lap: if it failsover and it get's anoter IP, my client... my dns resolving is not working anymore ?
[18:24] <patdk-lap> why would it get another ip?
[18:24] <YamakasY> patdk-lap: because I see that happening
[18:24] <patdk-lap> and why would it break dns?
[18:24] <patdk-lap> dhcp is dynamic
[18:24] <patdk-lap> you give your machine a random ip
[18:24] <YamakasY> no it's not, it's static if you set it
[18:24] <patdk-lap> if it is static, what does it matter at all?
[18:25] <patdk-lap> you really did not read anything I said
[18:25] <YamakasY> heh, server that get their IP from dhcp and switch IP, what do you think that will happen to dns resolving ?
[18:25] <patdk-lap> if it is istatic, they all know what it is, so why would it ever get a new ip?
[18:25] <patdk-lap> if it is dynamic, I said above, the dhcp master/slave will talk to each other
[18:25] <YamakasY> no it's only static in my leases file
[18:25] <patdk-lap> heh?
[18:25] <patdk-lap> it's static or dynamic
[18:25] <YamakasY> it's written to the leases
[18:25] <patdk-lap> leases file is leases, it has NOHTING to do with static or dynamic
[18:26] <patdk-lap> is the leases file a config?
[18:26] <YamakasY> whe you provision with something like foreman it does
[18:26] <patdk-lap> if not, it doesn't matter
[18:30] <YamakasY> patdk-lap: I'm confused
[18:31] <patdk-lap> yes, that was ovious
[18:32] <patdk-lap> cause your putting your assumptions on software, that does not work the way you think it does
[18:32] <YamakasY> patdk-lap: no its more how foreman handles it's DHCP leases... it's a way but dhcp failover itself works different :)
[18:32] <patdk-lap> well, I don't know what a foreman is
[18:33] <patdk-lap> or what it has to do with dhcp
[18:33] <patdk-lap> you ask about dhcp failover on ubuntu server
[18:33] <patdk-lap> ubuntu server uses isc dhcp server by default
[18:33] <patdk-lap> all infomation I gave is related to isc dhcp server
[18:33] <YamakasY> patdk-lap: it provisions dns, dhcp, tftp, puppet and provisions a server on a virtual or baremetal environment
[18:33] <patdk-lap> maybe try #foreman?
[18:33] <patdk-lap> this is #ubuntu-server though
[18:34] <YamakasY> patdk-lap: yeah but they mostly say, go to your distro or the package channel
[18:34] <patdk-lap> so they claim, foreman doesn't do it
[18:36] <YamakasY> patdk-lap: where do you see that ?
[18:36] <patdk-lap> they told you, they don't know, and don't care
[18:36] <patdk-lap> so it cannot be related to foreman
[18:36] <patdk-lap> therefor it works how your os does it
[18:36] <patdk-lap> how UBUNTU does it
[18:36] <patdk-lap> so you cannot come in here and say foreman does this or that
[18:36] <YamakasY> oh no they care but we are all checking new ways out and committing it in if needed
[18:36] <patdk-lap> if foreman people claim it doesn't
[18:36] <YamakasY> relax mate
[18:37] <patdk-lap> well, read up on isc dhcp
[18:37] <YamakasY> just collaborating
[18:37] <patdk-lap> it's all in the manual, whole section on it
[18:37] <YamakasY> yeah did but not about the leases file
[18:37] <patdk-lap> it's pretty easy and straight forward
[18:37] <patdk-lap> cause like I said, leases file has nothing to do with it
[18:38] <YamakasY> heh, sure it does
[18:38] <YamakasY> but in a different way
[18:38] <patdk-lap> what?
[18:38] <patdk-lap> what does leases file have to do with failover?
[18:38] <YamakasY> ok, end of discussion, you are not that friendly in words
[18:38] <YamakasY> as I said relax
[18:38] <patdk-lap> caus eyou keep counterdicting
[18:38] <patdk-lap> I said it doesn't have to do with failover, and you answer it does
[18:38] <patdk-lap> but you don't know why
[18:39] <YamakasY> have you ever checked a leasesfile between 2 failover systems, LB-ed or not ?
[18:39] <patdk-lap> it has nothing to do with failover
[18:39] <YamakasY> check that and come back
[18:39] <patdk-lap> leases file is a result
[18:39] <YamakasY> yes
[18:39] <patdk-lap> the infomation in it is the reault of the sync of failover
[18:39] <patdk-lap> leases has no place in failover though
[18:39] <YamakasY> no but I'm checking out a way it can
[18:39] <patdk-lap> it can what?
[18:40] <YamakasY> I'm thinking about a cold failover with rynsc otherwise
[18:40] <patdk-lap> why not use normal failover?
[18:40] <patdk-lap> failover that is cold, offline, is not a very useful failover
[18:40] <patdk-lap> cause when you need it, you cannot rsync
[18:41] <YamakasY> patdk-lap: because my information of my hosts is in my leases file and not in my config
[18:41] <YamakasY> as forman writes to the leases file
[18:41] <patdk-lap> directly writes to the leases file?
[18:41] <YamakasY> yap
[18:41] <YamakasY> why not ? you can
[18:41] <patdk-lap> ok, so it went way into unsupported territory
[18:41] <patdk-lap> you can do many things, doesn't mean they are supported :)
[18:41] <YamakasY> yeah, but isn't that with all software ?
[18:42] <patdk-lap> no
[18:42] <YamakasY> sure
[18:42] <patdk-lap> there is a api to use
[18:42] <YamakasY> sho supports isc when they stop it ? indeed no-one
[18:42] <YamakasY> so everyone can pickup
[18:42] <patdk-lap> not to use it and go around it, means what your doing could break at any time
[18:42] <YamakasY> it's like in insurance company ;)
[18:42] <patdk-lap> and it also causes you to be stick to specific software to do things
[18:42] <YamakasY> patdk-lap: hehe and you think a company as RedHat doesn't know how to handle that ?
[18:43] <patdk-lap> heh?
[18:43] <YamakasY> indeed, foreman is RedHar
[18:43] <patdk-lap> it would cost them money
[18:43] <YamakasY> *RedHat
[18:43] <YamakasY> huh ?
[18:43] <YamakasY> why ?
[18:43] <YamakasY> if they support it
[18:43] <YamakasY> you don't know how support works it seems
[18:43] <patdk-lap> someone has to fork dhcp, continue to support it, make sure that unsupported interface works, maintain security patches, ...
[18:43] <YamakasY> support is what you can handle as you know how it works and YOU can support your way
[18:44] <patdk-lap> if you did not do all that, and used a supported interface, the overhead is much much less
[18:44] <YamakasY> patdk-lap: yes and if I don't work today I might have more rest tomorrow :P
[18:44] <YamakasY> wtf, it's not an if/else
[18:46] <patdk-lap> for support?
[18:46] <patdk-lap> sure it is
[18:46] <patdk-lap> once you modify something, or don't do something the supported way, you end up supporting it
[18:46] <patdk-lap> this is a basic rule of bussiness
=== jelly-home is now known as jelly
[18:47] <patdk-lap> this is why there are so many cloud things
[18:47] <patdk-lap> offload the support issue with maintaining email, website, acocunting software,.... to whoever does it in the cloud
[18:47] <patdk-lap> or contract another company to do it
[18:48] <patdk-lap> if the company wanted to deal with supporting it, it would all be inhouse
[18:48] <YamakasY> patdk-lap: indeed that is where the company that suggests the modifications is supporting it
[18:48] <YamakasY> ever seen how github forks work ?
[18:48] <YamakasY> indeed like that
[18:48] <patdk-lap> once you fork, your stuck
[18:49] <patdk-lap> maintaining it forever, till you stop using it
[18:49] <YamakasY> yeah so ?
[18:49] <patdk-lap> unless you get your adjustments pushed upstream
[18:49] <YamakasY> if that makes your money, be happy
[18:49] <patdk-lap> sure
[18:49] <patdk-lap> but it if produces no money
[18:49] <patdk-lap> that is horrible way to do things
[18:49] <YamakasY> I'm not a as-if user
[18:49] <YamakasY> as-if people don't improve and are stuck
[18:49] <YamakasY> and actually don't know that much of their systems
[18:50] <an3k> party party. 7zip can't handle ~ at all, it creates a directory with that char ...
[18:58]  * YamakasY brings in the popcorn
[19:30]  * lordievader wonders if foreman writes to the leases file directly or through an api
[19:33] <lordievader> I guess [1] answers my question. [1] http://projects.theforeman.org/issues/748
=== Lcawte|Away is now known as Lcawte
=== Monthrect is now known as Piper-Off
[23:04] <grendal_prime> i got one werid issue
[23:05] <grendal_prime> setting up a rudundat router right...i have a standard ethernet device, and...a verizon jetpack device (usb connected)
[23:06] <grendal_prime> this router is an ubuntu server...and im using drsounds fault_tollerant_router  (ruby scripts basically)  Now i have tested this in virtual environments several times and...it all seems good.
[23:07] <grendal_prime> if i have one or the other connected...everything seems to work fine, but when i have them both connected, i have really werid dns issues.
[23:07] <grendal_prime> cant get to google....then all the sudden i can...
[23:08] <grendal_prime> very strange.
=== Lcawte is now known as Lcawte|Away