=== mikal_ is now known as mikal [02:41] I seem to be having an issue with isc-dhcp-server on Ubuntu 13.04, I notice in my logs that I am constatnyl getting this: " Can't create new lease file: Permission denied: 14 Time(s) " I tried troubleshooting, by tearing down apparmor, looking up the permissions, stopping isc-dhcp, chowning the permissions for the lease file to dhcpd:dhcpd, then starting the services again. They are then chown'ed to root:root [02:42] I can't figure out what the cause of this is, or how to permanently resolve it. === balloons_ is now known as balloons === mikehale_ is now known as mikehale === daker_ is now known as daker === irv_ is now known as irv [02:45] I did find in the startup script that at the end, it chowns them to root, but that wouldn't quite make sense because it would need to be dhcpd:dhcpd for the isc-dhcp server to modify the files. [02:45] But I didn't want to go about editing the startup script without making sure that would resolve the issue. === arrrghhhAWAY is now known as arrrghhh [03:07] hi, there was a power outage and now this ubuntu server doesn't connect to network [03:08] at startup it says "waiting up to 60 more seconds for network configuration..", the lights on NIC and ROUTER are blank [03:08] what can I do ? [03:10] Test a nother nic? [03:11] Set a static IP and see if you can connect to the network via that way? [03:12] GH0: I went into /etc/network/interfaces and removed everything except loopback as some people have suggested, it's booting up [03:12] yup, see some light on the router [03:13] No idea after that, that would have been my basic trouble shooting steps, then I would have looked online, then I would have come here. [03:17] ok so with only loopback, the light on router is blinking green, but as soon as I do ifconfig eth0 up 192.168.2.25 [03:17] the light goes dead [03:19] what log file can I look at to see errors? [03:21] dmesg shows "eth0: link is not ready" [03:25] I'll download knoppix and see if the problem is with ubuntu or NIC [03:25] so silent here... [03:29] Nicolas_Leonidas, unfortunately, it is. I was having issues with isc-dhcp-server and am unable to figure them out so I am waiting. [03:33] Nicolas_Leonidas: what does dmesg and syslog say about what happened? [03:36] GH0: dhcpd should be running as root [03:36] i don't have a 13.x system at my disposal [03:37] What permissions are needed for the dhcpd.leases file? [03:37] Maybe my permissions are just broken, but, rw-rw---- would seem appropriate to me. [03:37] -rw-r--r-- 1 root root 5076 Jul 10 20:16 /var/lib/dhcp/dhcpd.leases [03:38] tedski: link not ready [03:38] $ ls -ld /var/lib/dhcp [03:38] drwxr-xr-x 2 root root 4096 Jul 10 19:59 /var/lib/dhcp [03:38] :/ Those are the same permissions I have. [03:38] Nicolas_Leonidas: are you sure the power outage did not cause a hardware failure? [03:39] tedski: no I'm not sure I'm downloading knopiix in the setup you can open a browser, that's a good idea right? [03:39] Nicolas_Leonidas: that's a start [03:40] Nicolas_Leonidas: i also blow out /etc/udev/rules.d/*persistent-net.rules [03:40] Nicolas_Leonidas: that will ensure udev hasn't mixed up your hardware ordering [03:42] GH0: what does apparmor say when you put it in complain mode? [03:44] GH0: also, what are ther permissions on dhcpd.leases~? [03:44] (note the tilde) [03:45] The permissions are the same as the non-tilde file. However, how do you put apparmor into complain mode? aa-complain doesn't seem to work, though that is probably a really old command. [03:46] https://help.ubuntu.com/community/AppArmor#Put_a_profile_in_complain_mode [03:46] aa-complain is the correct usage [03:47] sudo: aa-complain: command not found [03:48] Yeah, that is what I found... [03:48] I would figure that command would be installed with the apparmor package. [03:50] i'm at a talk right now, so i actually don't have access to any ubuntu boxen :( [03:50] so, wish i could get help [03:50] my debian boxes don't run apparmor [03:50] i'll be back later [04:00] sarnold: ^^^ === Guest37357 is now known as G [07:17] failed to detect os error while installing virtualmin on 12.04 server [08:15] Daviey: yep === Adri2000_ is now known as Adri2000 [08:38] yolanda, lemme pull that merge and take a look [08:38] now finding that error: undefined reference to `libecap::Area::toString() const' [08:42] jamespage, problem with a diff of debian/control [08:48] yolanda, did you managed to drop a dependency? [08:49] jamespage, i did an incorrect merge in control and dropped a dependency, yes [08:50] yolanda, OK _ just reproducing your issue above [08:51] hi, how am i able to force NTP to update my local system clock ? [08:51] the offset is 264seconds, i need it to be 0 === Guest30554 is now known as g0tcha [09:00] yolanda, OK - I don't see that problem - are you sure the control line for BD's is OK now? [09:00] double checked now === Duologic_ is now known as Duologic [09:02] yolanda, hmm - libecap2 is in universe [09:03] not in main, so is not ok to add as a build-dependency? === Jeeves__ is now known as Jeeves_ [09:08] yolanda, well it's been added since the last merge - it looks sensible so we should do a MIR for it [09:09] yolanda, I'd include it in the merge and then file the MIR for libecap [09:09] yolanda, have you done any MIR's yet? [09:09] jamespage, never [09:09] yolanda, https://wiki.ubuntu.com/MainInclusionProcess [09:09] yolanda, its actually quite a simple one so a good one to start with === ivoks_ is now known as ivoks [09:10] jamespage, ok [09:10] no extra dependencies or suchlike [09:10] libecap2 and libecap2-dev [09:10] yolanda, but do the merge first and then worry about the MIR [09:10] ok [09:10] yolanda, yeah - source is libecap [09:10] i submitted a merge request for nagios3 yesterday [09:10] just filed a bug and sent the debdiff [09:13] yolanda, for squid3 there are a couple of other bits that can be fixed [09:14] the transitional squid and squid-common packages can be dropped now [09:14] i did it [09:14] and refreshed patches [09:14] yolanda, did you drop two patches? [09:14] 98 and 99? [09:14] the 98 [09:14] refreshed the 99 [09:15] there are two 99 anyway, the ssl-cert-snakeoil and the 99-lp [09:16] i found that there is a line that applies in 99-lp, the -667,7 one [09:17] yolanda, okay - good [09:17] (I did not look that hard) [09:18] jamespage, i preserved the apparmor stuff [09:18] good [09:18] and autopkgtests, i think it's all our delta [09:20] yolanda, re 99-lp - that patch came from upstream so I don't think its still required [09:20] http://bazaar.launchpad.net/~squid/squid/trunk/revision/12473 [09:21] the patch is slightly different from upstream, but it makes sense to drop it, yes [09:22] yolanda, it will be because it was backported [09:22] upstream move on afterall [09:23] ok === andol_ is now known as andol [10:15] dosaboy: Sorry, just saw your response. bug 1199037, are you working on it? [10:15] Launchpad bug 1199037 in python-eventlet "backport eventlet exception context fix" [Undecided,New] https://launchpad.net/bugs/1199037 === jtv1 is now known as jtv [10:25] Daviey: yes [10:25] well not right now but I aim to have it done soon [10:26] 0.13 is currenlty being packaged for Ubuntu as well [10:26] dosaboy: OK, thanks [10:26] dosaboy: by who? [10:26] zul [10:27] not sure which archive it is aimed at though [10:27] dosaboy / zul: Why are we jumping ahead on this one? [10:27] so I am gonna backport that fix to raring/cloud archive [10:27] dosaboy: Nothing should be targeted at the cloud-archive that isn't in the backing primary archive [10:28] Daviey: this backport is aimed at primary and cloud archive [10:29] the 0.13 release I am not sure, you'll have to ask zul [10:29] dosaboy: ok, thanks [10:32] Daviey: just noticed I had not assigned that bug ;) [10:32] sorted now [10:33] ta! === rvba` is now known as rvba [11:15] To develop one tcp server ,which design is better ? multithread and multiprocess [11:36] adam_g, do we have a juju-deployer compatible with juju-core yet? you mentioned something yesterday? [12:21] jamespage, is ok to grab erlang for a merge? [12:21] yolanda, already did it [12:22] ok, i'm looking at robies html [12:22] maybe it's outdate [12:22] outdated [12:22] libnss-ldap is ok then? === mic__ is now known as mic_ [12:36] yolanda, yeah - it does not autorefresh I think [12:36] yolanda, libnss-ldap - go for it === cmagina-away is now known as cmagina [12:44] ok [12:45] some conflicts with patches mostly [12:46] good lawd the internet is a wasteland === exekias_ is now known as exekias [13:45] Anyone willing to help a newb manually upgrade certain packages? I have myself in a wicked tangle right now. I installed a package with dpkg (qemu-kvm from cloud archive), but I need to install some dependencies (like libbluetooth3) but it won't let me install that because it compains qemu-kvm has unmet dependencies. [13:47] what does 'apt-get -f install' say? [13:48] Jeeves_: it tries to remove a bunch of packages that I really do not removed. [13:49] Jeeves_: http://pastie.org/8131248 [13:50] namely nova-compute and nova-compute-kvm (qemu-kvm and qemu-system-x86 I installed manually with dpkg -i) [13:52] Is there an easy way to undo the manual installs that I did? I am am so new to debian/ubuntu package management. Sorry :/ [13:52] Chocobo, usually i use apt-get so it deals with dependencies itslef. [13:53] like, once the cloud archive ppa is added to the system, apt-get update, and then apt-get install whetever i need from there [13:54] Chocobo, if the ppa is correctly added, what happen if you apt-get update, and apt-get install qemu-kvm ? [13:55] melmoth: I would love to do that, but I am using Grizzly which does not include a newer version of qemu (>=1.3.0) which is required for 0-downtime snapshotting. I am trying to install a newer version manually from the cloud archive. [13:55] i dont get the "from the cloud archive" bit. If the package you want comes from the cloud archive ppa, you shoul dbe able to use apt-get [13:56] if it does not, then, yeah, you ll have to install all dependencies first before dpkg -i works. [13:57] melmoth: Grizzly: http://ubuntu-cloud.archive.canonical.com/dists/precise-updates/grizzly/main/binary-amd64/Packages and Havana: http://ubuntu-cloud.archive.canonical.com/dists/precise-updates/havana/main/binary-amd64/Packages [13:57] If you do a search for "Package: qemu" on those sites you will see that hava includes the newer version of qemu but grizzly does not. [13:59] melmoth: any idea how to satisfy the dependencies when apt-get refuses to work because of missing dependencies? it seems like "dpkg -i" installs the package even if there are missing dependencies which causes problems if you try to use apt-get? [13:59] Chocobo, what i am about to propose may be a bad idea, but hey.. what about 1) adding the grizzly ppa, 2) install qemu-kvm from there (with dependencies) [13:59] then removing the havana ppa, adding the grizzly one, and apt-get install the stuff you need from there [14:00] i dont know if dpkg -i install things breaking dependencies on top of my head, but i hope it would not do that :) [14:00] i meant, 1) adding the havana ppa [14:01] melmoth, Chocobo: please hold [14:01] htmmm, I could try that. Not a bad idea. [14:01] jamespage: ok. [14:01] Chocobo, we only backported qemu to havana [14:01] so grizzly uses 12.04 version [14:01] * jamespage reads backscroll [14:03] Chocobo, ah - I see - so you have picked the qemu from the havana cloud archive - but you don't have it installed as a specific source right? [14:03] jamespage: I really need qemu >= 1.3.0 in Grizzly... I need to find a way to manually install it. Correct, I do not hava Havana set as a source (because we are using Grizzly) [14:04] Chocobo, OK - this is not officially supported in any way [14:04] but you could add the havana CA to your sources and then use pinning to ensure if does not get picked by default [14:04] so you can grab qemu from havana and openstack from grizzly if that makes sense [14:05] jamespage: I think so. I am not sure about the pinning part. [14:05] Chocobo, http://jaqque.sbih.org/kplug/apt-pinning.html [14:05] Chocobo, if you follow that but use the cloud-archive repos [14:05] and ping havana lower than grizzly [14:06] you should be able to "apt-get install qemu-kvm/havana" for example [14:06] jamespage: and that will handle dependencies? [14:06] * jamespage crosses his fingers [14:06] I think so yes [14:09] so.. I can't use "release" as my pin... because it looks like the release for cloud archive is "main" is there a pin for "repo" or something similar so I can ping precise-updates/havana lower than precise-updates/folsom? [14:10] Hold on, I may have found the man page on pinning [14:11] anyone got an executable tcping http://www.linuxco.de/tcping/tcping-1.3.5/ [14:11] ? [14:11] i can't be bother to install dev tools just for this [14:13] Chocobo, Pin: release c=precise-updates/havana [14:13] I think [14:15] jamespage: and a Pin-Priority of probably.. 0 < P <=100 [14:15] Chocobo, something like that [14:15] you can alway check things by doing 'apt-cache policy qemu-kvm' [14:15] it will tell you the orders of install by default [14:16] Chocobo: When you have the formula, can you pastebin it please? [14:19] I will. Hrmmm, I hope the troubles I caused earlier is not going to get in my way. [14:20] jamespage: E: Release 'havana' for 'qemu-kvm' was not found [14:23] Chocobo, http://pastebin.com/H3MfSJ12 [14:24] Chocobo, with that approach you don't even need the /havana [14:25] jamespage: I think I need to find a way to undo the trouble I caused with dpkg :( [14:26] nevermind. i should use nc instead [14:27] Wow, I am terrible at Ubuntu package management. I feel like I am all thumbs. [14:28] and I keep going in circles. Ok, so lets say I installed a package with "dpkg -i" and it had unmet dependencies... how can I undo that? [14:28] undo it or fix it? [14:28] usually you can sudo apt-get -f install afterwards [14:29] and it will fetch the missing deps and then resolve itself [14:29] zul, actually qemu-kvm is uninstallable from the havana cloud archive right now [14:29] jcastro: I can't do that because it want's to uninstall some of my openstack components. What I really want to do is replace the manually installed version with the version in repository. [14:30] jamespage: uh? [14:30] zul, vgabios and seabios [14:30] qemu-system-x86 : Depends: vgabios (>= 0.6c-3~) but 0.6c-2ubuntu3 is to be installed [14:30] Depends: seabios (>= 1.7.2-2~) but it is not going to be installed [14:31] So this won't work even if I fix my problem from dpkg? [14:31] Chocobo, nope [14:31] Well crap [14:31] havana cloud archive is still work in progress [14:32] consider it as stable as the development release day-to-day [14:32] I am going to have to this from source. :/ [14:32] How can I undo the changes I made with dpkg? I want to go back to the version from the cloud-archive. [14:33] jamespage: ohhhh sshhi.. [14:36] These are the changes I made: http://pastie.org/pastes/8131364/text [14:39] There doesn't seem to be an option to ignore dependencies in apt-get [14:51] Is there anyone that might be able to help me? I am really stuck here. === shirgall_ is now known as shirgall [15:00] Gosh, I think I got it... I needed --force-depends [15:03] smoser: do we have docs on simplesteams anywhere? [15:03] streams even [15:05] jcastro, not very good ones. but there is some in lp:simplestreams doc/ [15:05] jamespage: http://people.canonical.com/~chucks/ca/ [15:06] smoser: I'd like to put that on developer.ubuntu.com if you don't mind === arosales_ is now known as arosales [15:09] jdstrand: Just curious if any luck in reproducing bug #1197484 [15:09] Launchpad bug 1197484 in isc-dhcp "Connection requests to saucy server VMs from a hosts fail after fresh VM installs" [High,New] https://launchpad.net/bugs/1197484 [15:11] how can I respond to a Broadcast message when I'm logged in? [15:12] psivaa: no, but I did see it on a lease renewal, so I have something to chase after [15:13] ndee: depends on what you mean by a broadcast message. Are you looking for wall(1) or write(1)? [15:13] rbasak, it was "wall" I think :) thanks [15:13] jdstrand: ack [15:19] does anyone know why, when I connect an iscsi target to my ubuntu box it sends 12 megabits of data, from my system to the iscsi host continously for no reason whatsoever? === kirkland` is now known as kirkland [15:26] it has data it wants to save? [15:27] oviously, your guess of, no reason whatsoever, is not accurate === Jikan is now known as Jikai === Jikai is now known as Jikan [15:29] zul, looks OK - do they build? [15:29] with precise havana [15:29] jamespage: yeah lemme double check === guampa is now known as dir [15:34] jamespage: yep [15:34] zul, +1 [15:36] jamespage: can we just put juju-core in saucy today? [15:36] mgz, otp - free in a bit [15:36] sure, poke me when done [15:49] Where can I find the source for packages? [15:49] (from cloud-archive) [16:09] mgz, OK - done [16:10] mgz, there are some challenges with the tarball up on launchpad.net - namely the top-level source folder is missing [16:10] so I could not load it directly into the packaging branch like we did last time [16:11] mgz, I think dcheney proposed a branch - but it would be great to cut a new tarball if possible [16:11] but we can do that step manually no? ( should have caught it in review of the script, sorry) [16:12] mgz, well we can yes [16:13] basically, we need a 1.11.3 pretty soon anyway to fix a few crucial upgrade bugs [16:14] gah, network connection being unreliable [16:15] mgz, OK - I'll manually repack the tarball for now on the assumption that next release it will be fixed up [16:15] so juju-core-XX.XX.XX/src/ [16:15] OK [16:15] so, we could reroll the tarball tomorrow, but seems like that step is just one of several niggles we can shake out before a cleaner release either tomorrow or early next week [16:15] yeah, that sounds good [16:16] Anyone know of a web based replacement to onenote that we can host on our servers? [16:17] mgz, you will want to be aware of bug 1200255 [16:17] Launchpad bug 1200255 in golang "go get ... fails with SIGILL on armhf" [Undecided,New] https://launchpad.net/bugs/1200255 [16:17] mgz, I'm intending on enabling the armhf arch for this next upload.... [16:19] enabling arm sounds good, if if blows up, we need to know [16:19] rbasak, bug 1200255 [16:19] Launchpad bug 1200255 in golang "go get ... fails with SIGILL on armhf" [Undecided,New] https://launchpad.net/bugs/1200255 [16:20] that bug is new to me... [16:20] I'm not using golang 1.1.1 on my box of course [16:21] mgz, uploading right now [16:21] mgz, thats just on ARM [16:21] ah - of course [16:23] I've been living with cgo being borked on my arm machine, should try the 1.1.1 from ppa [16:23] mgz, please do [16:24] hey, im wondering if someone might be able to help me solve some openssl issues on 12.04 vs 10.04: https://gist.github.com/NoiseEee/b99e4994329ddab53fa9 [16:29] mgz, OK uploaded [16:29] ace. [16:30] what's your packaging branch, so I can pull? === mist is now known as Guest918 === Guest918 is now known as mist [16:38] jamespage, i have a branch i was workin on that works now with py-juju. need to cleanup and make surei didnt break the juju-core interface === genii_ is now known as genii [16:41] mgz, it will be ubuntu:juju-core once the upload is built [16:44] is there a reason i might have a /srv/cvs folder on a serveR? i've not installed CVS === matsubara is now known as matsubara-lunch [16:57] GH0,tedski,ScottK, aa-unconfined might not be installed because the apparmor-utils package drags in another 40+ megabytes of packages via hilarious dependencies. :( [16:58] alright, is there a specific package to run it? [17:01] Hi! Does anyone have experience with ircd-hybrid and hybserv? I'm having some trouble getting NickServ and the like available. I can see them when I log into my server from itself, but not from the outside. [17:06] :( [17:07] Oh well. Thanks anyways guys! [17:08] just for like...being on irc. === Guest36906 is now known as med_ [17:12] zul, are these issues on saucy or precise? https://bugs.launchpad.net/nova/+bug/1200231 [17:12] Launchpad bug 1200231 in nova "Nova test suite breakage." [High,New] === marass is now known as mcas === matsubara-lunch is now known as matsubara [18:19] adam_g: saucy [18:19] zul, have you tried with the newer sqlalchemy? === Aww_ is now known as Aww [19:24] TheSov: that sounds like a bug. Can you check for an existing bug, and if there isn't one then make sure that you can reproduce the problem on a freshly installed system (and a fresh iSCSI target) and then file a bug with instructions to reproduce the problem? [19:26] Hey folks, anyone know if there is any local disk caching that happens with cifs shares? I've got a file server that claims it is nearly out of disk space, but appears to be counting a mounted cifs share against that total. 0.o [19:30] As far as I can tell, I've got a 300gb dir which puts me at about 45% capacity. That directory is being rsynced over a cifs share to a backup file server. It appears that location I mounted the cifs share is counting against the total root file system capacity, so my server thinks it is nearing 100%. It actually denied a write to another server, saying that the disk was full. === Nolar2_ is now known as Nolar2 [19:31] Any guesses or smacks with the clue bat would be appreciated... my googling isn't turning up anything helpful. === LordOfTime is now known as TheLordOfTime [19:32] psyferre: can you paste df output? [19:32] psyferre: have you accidentally rsynced at a time when the cifs share was not mounted? When mounted again, it will be hidden. [19:33] rbasak: you're a genius. I bet that's it. [19:34] Because I've never ever forgotten to mount something and then accidentally filled the root fs up. Oh no. :-P === Guest5603 is now known as guntbert [19:35] Hi all I'm trying to export a nfs share with no_root_squash but when I mount it on my client and write a file with root this file is squashed nobody:nogroup. What I'm forgetting ? === guntbert is now known as Guest14360 [19:35] ( sorry for my bad eng ) [19:35] rsync over a cifs share though? I never trusted that, especially for backups. Have you considered running an rsync daemon on the backup server? cwrsync has worked well for me for Windows servers in the past. [19:37] rbasak: I think I tried that and had a bunch of trouble getting it setup. Then I found that I could mount the share with _netdev and it would *supposedly* remount if it lost connection... [19:38] I'll go back to the drawing board and find a better way to sync. === Guest14360 is now known as guntbert === masACC is now known as maswan === genii_ is now known as genii === acrocity_ is now known as acrocity === LordOfTime is now known as LordOfTime|EC2 [20:23] adam_g: not yet [20:24] adam_g: its still stuck in -proposed [20:24] zul, can you pip install it and try the test suite ? [20:24] adam_g: sure === Gasseus is now known as Rallias === _thumper_ is now known as thumper [22:05] Hello! Could someone tell me typical values (watt) what a server consumes normally? [22:05] [New hardware, about 900eu / 1000 usd] [22:12] grex25: in the wild-guess territory, I think 500 watts at boot, 350 watts at moderate load, 125 watts at light load [22:13] should be very energy efficient, what are "good" values? 80? [22:14] Normally there it should not have a to high load, but raspberry is to slow [22:17] grex25: I think laptops under full-tilt will get closer to 80 or 100 watts, they'd still be way faster than a raspberry [22:17] grex25: (I've got a pandaboard that I think is about 5 watts; it's sigifnicantly faster than the raspberry, but not exactly fast.) [22:18] sarnold: Now its a intel atom setup, 50W, but I thought there would be something better now (2/3 years old hardware) [22:20] grex25: intel's new haswell chips come in a variety of top-power-draw .. server editions can go for ~100 watts, laptops can be quite low... but of course that's just one component of the whole [22:20] sarnold: but panda/beagle/raspb are very nice [22:20] grex25: I get 10MB/s IO to my pandaboard's CF. take that into consideration.. :) [22:21] sarnold: okay, than i have to wait a little bit [22:21] grex25: I'm not saying it isn't nice :) it just isn't suitable for all server tasks. irc and torrent host? awesome. hehe. === cmagina is now known as cmagina-away [23:40] howdy [23:40] i used to be able to follow this guide, and it did work for some systems in the past, but recently, i have tried this on two more new installations and i always hang at updaing yasm to 2.0 [23:40] err 1.2.0 [23:40] https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide [23:41] i installed a fresh 12.04LTS, and followed the steps. this did work for me in the past, no problems, just copy and paste each block [23:42] now, i do the yasm compile instructions line by line, they all seem to complete, but when i do yasm --version i still get the 'old' one. [23:43] nvm [23:43] anepanaliptos: if you installed yasml from the archive, it is liable to be in your PATH before the version that you just compiled and installed.. [23:43] yeap. and that's exactly what it ius. just thought about it. [23:43] id 10 t error. [23:44] anepanaliptos: and if your new one _should_ be in your PATH earlier, perhaps you have to tell your shell to delete the hashed location of the executable, hash -d ought to do it.. [23:44] anepanaliptos: woot. :) [23:45] root@tesla:~/ffmpeg_sources# which yasm [23:45] /usr/local/bin/yasm [23:45] root@tesla:~/ffmpeg_sources# yasm [23:45] -bash: /usr/bin/yasm: No such file or directory [23:45] root@tesla:~/ffmpeg_sources# [23:45] just fixed it with a symlink [23:46] hash -d yasm also should have fixed that, fwiw [23:46] not sure if there is a 'proper' way [23:46] (well i did do the hash -d before i ran that command) [23:47] oh? hrm. [23:47] but no worries. the major part is fixed. [23:47] maybe your /usr/bin/yasm was missing a library. that's not as common as it used to be.. [23:48] ah. I was wrong. hash -d doesn't delete everything. bah. :)