IRC logs for #baserock for Wednesday, 2014-11-05

*** genii [~quassel@ubuntu/member/genii] has quit [Remote host closed the connection]00:19
* persia thinks the needs of web developers differ from the needs of GUI developers differ from the needs of service developers differ from the needs of tool developers, leading to "devel-system" being nomenclature that encourages discussions of paint colour01:06
*** cosm [] has quit [Ping timeout: 264 seconds]01:43
*** cosm [] has joined #baserock01:45
*** petefoth_ [] has joined #baserock01:50
*** petefoth [] has quit [Ping timeout: 260 seconds]01:51
petefoth_ is now known as petefoth01:51
*** sambishop [] has quit [Ping timeout: 244 seconds]02:47
*** zoli_ [] has joined #baserock04:22
*** zoli_ [] has quit [Changing host]04:22
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock04:22
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]06:03
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock06:04
*** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 245 seconds]06:09
*** zoli_ [] has joined #baserock06:38
*** zoli_ [] has quit [Changing host]06:38
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock06:38
paulsher1oodin other news, the databases stratum builds fine on jetson07:10
*** mike [] has joined #baserock07:21
mike is now known as Guest3970307:21
*** dutch [] has joined #baserock07:36
* persia likes infinite scrollback, but trying to download 6 months of messages looking for something specific 200 lines at a time makes me very happy to have logging (can we take it out of test?)08:02
* persia finds a discussion with paulsher1ood and pedroalvarez about factory upgrades being broken from August08:12
* persia thanks ssam2 for having provided excellent instructions08:13
pedroalvarezTime to fix them! 08:22
pedroalvarezRegarding to improving the logs, I'm researching in my spare time about it. 08:24
persiaThe start is lovely.  For all that things could be better, I believe I now have a better interface for #baserock logs from now.08:30
persiaAnd as long as we have the data, we can get everything from when logging started into the right format08:31
persia(and for that matter, once we have a final format, several of us can probably donate logs for rationalisation into that format to reach further back)08:31
*** tiagogomes [] has joined #baserock08:37
paulsher1oodcool :)08:47
*** mariaderidder [] has joined #baserock08:57
Kinnisonjjardon: gcc.git took me 30 minutes to clone08:58
Kinnisonjjardon: I'm loathe to add it unless you really want to be working from git08:58
Kinnisonjjardon: I'd rather you just updated the tarball lorry to pull a newer version08:58
Kinnisonjjardon: Also it's a 1.5G git repo, which is a little on the large side08:58
persiaWhat is the slow thing here?09:01
Kinnisonit's a GNU project09:01
* persia really doesn't like tarballs, because there is not good automation to update09:02
KinnisonGNU projects are, almost universally, the worst thing for git09:02
Kinnisonesp. long lived ones09:02
persiaThat's not enough.  Is it that the server is slow, is it that the latency is high, is it conversions, etc.?09:02
KinnisonWe're not converting, their server is reasonably fast, as is my laptop.  of the 90m, nearly 45m was CPU time on my laptop09:02
KinnisonGNU projects tickle a pathological behaviour in Git -- specifically with their insistence on a reverse-order 'ChangeLog'09:03
Kinnisonevery commit (nearly) alters a file by making it longer at the top -- Git is "not good" (tm) at that09:03
Kinnisonit makes the delta compression stuff for packs take aaaaaages09:04
persiaIsn't the pack delta algorithm selectable?  Are they all horrid?09:05
KinnisonNo idea if it's selectable or not, or how you would, and I've not experimented beyond trying to clone from gcc.git where one would assume they had selected an algorithm which was the best option for their repo09:06
persiaIs it faster with -a-d?09:08
paulsher1oodcould we somehow .gitignore their changelog? :)09:09
persiaMy apologies: that doesn't actually work with clone.09:09
persiapaulsher1ood: That wouldn't cause the set of objects on their server to be different, and partial clones are annoying09:10
Kinnisonpaulsher1ood: no, richard_maw and I had a brief chat about filter-branch becoming part of lorry, but that makes contribution back hard09:10
* Kinnison just doesn't want to add a half-hour clone time to anyone building gcc09:10
Kinnisonit's bad enough with linuc09:10
paulsher1oodah, hold on. 09:11
Kinnisonpersia: If you want to +1 the gcc lorry, I'll not stop you, I intend to -0 it09:11
Kinnison(I don't like it, but not strongly enough to veto)09:11
persiaPeople building gcc don't need everything: a shallow clone ought be enough09:11
paulsher1oodfor someone who *wants* to build gcc (eg will fiddle with it themselves)... the clone time seems unavoidable09:11
paulsher1oodfor others, if we have continuous-everything, they should get a cached gcc anyway?09:12
Kinnisonpaulsher1ood: and someone who cares not a whit for gcc but needed to alter fhs-dirs?09:12
persiaI want to solve the problem.  I don't actually care about the gcc lorry, and I don't think random votes on it help other than postponing the issue.09:12
Kinnisonpersia: there's tradeoffs everywhere09:12
paulsher1oodsorry, for mere mortals, what's fhs-dirs got to do with it?09:12
persiaI'd like to see if a different pack algorithm would make it faster, although repacking for lorry can be expensive.09:12
Kinnisonpersia: a 'shallow clone' doesn't help the next time you need to build a different SHA09:12
Kinnisonpaulsher1ood: it's lower in the stack09:12
persiaKinnison: Depends which SHA, and how shallow :)09:12
KinnisonIIRC git can't pull updates into a shallow clone without deepening it, no?09:13
paulsher1oodcan we tweak our stack? pretty sure we haven't got all the ordering fully optimised?09:13
persiaGrr.  Seems git doesn't do what I want.  If I want a fast pack, I need to have the fast pack on the server already, rather than there being ways I can ask the server to give me a fast pack, or somehow magically receive a different pack than the server provides.09:14
persiaBut even if upstream made a perfect pack, it would quickly go bad unless every person who pushes uses a pack algorithm that works well with add-text-to-massive-file-at-the-top09:14
* paulsher1ood notes that the 'failed to find tarball message' we killed last week may indicate that we used to have something to mitigate part of this, that may no longer be working09:15
KinnisonWe do have something which helps to mitigate the initial clone, but with a GNUish set of content, any time git decides to trigger a gc we're likely waiting for 45m to an hour09:16
Kinnisonand the tarball thing works09:16
petefothI believe that we just killed the error message, and we still lookf firts for a tarball and only go for git if we can’t find one (but I could easily be wrong about that)09:16
Kinnisonpetefoth: You're not wrong09:17
Kinnisonand we didn't kill an error message, we reworded an informational/warning09:17
persiaI guess what I want from tarball is a richer syntax, like that for watch files, that would generate a git repo with one commit per release, or similar09:17
Kinnisonpersia: to my mind, that's for something like (but not exactly like) firehose09:17
persiaIdeally a model that would add tags to the repo for every release of every branch of gcc, so one would swap versions and microreleases by changing shas.09:18
persiaKinnison: I consider that *input* for something like firehos09:18
Kinnisonpersia: our tarball import supports changing the tarball and thusly getting a new sha without losing the old one09:18
persiaI want that in *lorry*, because I want to get new commits when upstream releases new tarballs.09:18
Kinnisonpersia: yes, clearly the watch file is input09:18
paulsher1oodKinnison: if it works, why would i ever have seen that message?09:18
Kinnisonpaulsher1ood: because lorry generates the tarballs, so you don't get them for stuff which is inside your trove prefix09:19
persiaTo be clearer, I want the watch file as input to lorry, and lorry as input to integration automation09:19
Kinnisonpersia: I suppose we could extend lorry's tarfile support to have watch input09:19
paulsher1oodKinnison: for gbo, which tarballs should i expect it to find?09:19
Kinnisonpaulsher1ood: anything in upstream:09:19
* paulsher1ood believes he has seen that message often, for upstream:09:20
persiaKinnison: That's my thought of the least-surprise solution.  Not that it isn't a bunch of work :)09:20
* paulsher1ood can be wrong of course09:20
Kinnisonpersia: I suggest you put a suggestion on the ML -- perhaps someone will pick it up.  If we had a storyboard instance we could put it in there :-(09:20
Kinnisonpaulsher1ood: If your trove configurations were wrong you may have, but it usually works and works well09:21
Kinnisonpaulsher1ood: There's some work we could do to reduce the brittleness of the configuration09:21
pedroalvarezI believe that the message only happened with our repos09:21
persiaKinnison: Heh.  I'll refine it a bit and do that indeed, now that I have more confidence it isn't entirely insane.09:21
paulsher1oodchecking logs, i see most of my examples relate to the baserock-clone at datactentred09:26
KinnisonThen your trove configuration is probably incorrect09:26
Kinnisonyou need to ensure that the trove hostname and id match properly at both ends of the system (trove and morph) otherwise it can't find the tarball files09:27
KinnisonThis is the brittleness I said we could do some work to reduce09:27
paulsher1oodbut i can find one for g.b.o i think:
paulsher1oodtrove and morph? all i do is fix /src/morph.conf - is there something else?09:28
KinnisonIt depends on how lorry etc is configured on the trove09:28
*** jonathanmaw [] has joined #baserock09:33
ratmice________fwiw, from the perspective of someone bootstrapping non-linux stuff off of baserock, we have quite a few patches to gcc, and a) making sure these don't break linux b) maintaining 'em is very helpful OTOH it is a complete hog so i understand, but i was a bit dissapointed to be honest to see the tarball import09:33
Kinnisonratmice________: useful perspective, ta09:37
* Kinnison wonders if he should get some shears out to trim ratmice's tail09:38
ratmice________they did that to my friend uma :(09:40
* Kinnison promises these'll be clean ones, no rust or tetanus in sight09:42
*** dutch [] has quit [Quit: Quit]10:10
*** ssam2 [] has joined #baserock10:21
Mode #baserock +v ssam2 by ChanServ10:21
*** perryl [] has quit [Remote host closed the connection]10:22
*** perryl [] has joined #baserock10:27
*** zoli__ [] has joined #baserock10:42
*** zoli_ [~zoli_@linaro/zoli] has quit [Read error: Connection reset by peer]10:42
pedroalvarezfranred, ssam2: what do you think about and calling it
franredpedroalvarez, I like it11:09
pedroalvarezthe good thing about this piece of infra, is that we don't need to back it up :)11:11
radiofreehi, regarding the issue with upgrading on newever jetson kernels, using ROOT_DEVICE would solve the problem (at least on the jetson)11:12
radiofreei was thinking of sending a patch to use ROOT_DEVICE (if available11:12
radiofreeif not, do the same thing - read /proc/mounts11:13
pedroalvarezradiofree: how close are we to have our devel system with a newer kernel?11:13
pedroalvarezthat also would solve the issue :)11:13
radiofreepedroalvarez: no it wouldn't?11:13
radiofreethese newever kernels seem to be the cause of the problem11:14
pedroalvarezthen, I'm not sure about what issue are we talking about, sorry :/11:14
radiofree/proc/mounts lists /dev/root as /, bbut /dev/root doesn't exist11:14
radiofreeso upgrade fails trying to mount that11:14
radiofreehowever, for jetson upgrades/deploys you always need to set the ROOT_DEVICE anyway11:14
pedroalvarezradiofree: I believe that i fixed the need of setting ROOT_DEVICE11:15
pedroalvarezthe issue was fstab with /dev/sda, wasn't it?11:16
radiofreeand the kernel cmdline11:16
radiofree"root=/dev/sda" previously11:17
pedroalvarezradiofree: this one fixed the fstab one:
radiofreeoh right11:19
radiofreei believe that *broke* upgrading11:19
franredwhen we update morph should we update strata/morph-utils.morph and strta/cross-bootstrap.morph's sha1 ?11:19
richard_mawfor anything other than changes to the test-suite, yes11:20
radiofreepedroalvarez: now, on a jetson, it can't get the uuid from /dev/root, so it fails11:20
ssam2pedroalvarez: makes me a bit sad that it requires javascript, but better than nothing11:20
ssam2and good that it can be used from the console like !11:21
pedroalvarezssam2: yup! we can update to write some instructions :)11:22
pedroalvarezradiofree: is that kernel stable?11:22
radiofreeyes it's pretty stable i've been using it for a week now11:22
pedroalvarezradiofree: I believe that that is a bug in that kernel :/11:23
jmacsIs the extension auto-detected?11:24
radiofreepedroalvarez: do you have an old jetson devel board around?11:25
radiofreewith 3.10 running on it11:25
radiofreethat has a /dev/root i suppose11:26
pedroalvarezjmacs: the colours of the paste depends on the extension you use in the url. 11:26
jmacsOh, I see11:26
radiofreebtw my laptop doesn't have a dev/root11:26
radiofreeso i'd hardly call it a kernel bug11:26
jmacsSo is the same paste with different highlighting. Cool.11:27
pedroalvarezjmacs: also:
radiofreedo you have a /dev/root?11:32
radiofreeit actually get's the correct /dev/mmcblk0p1 device from proc/mounts there11:33
radiofreeon this kernel - /dev/root / btrfs rw,noatime,ssd,space_cache 0 011:34
petefothKinnison: did you have time to look at the latest version of the glossary that I pushed yesterday?11:36
KinnisonI ran out of brain11:36
KinnisonAnd this morning I forgot11:36
* Kinnison is a bad man, and he should feel bad11:37
pedroalvarezregarding this comment I made yesterday: <pedroalvarez> as a side note, I think cross-bootstrap is broken11:37
pedroalvarezplease,  ignore it, it was a full-disk issue.11:37
* Kinnison shall look now11:37
petefothKinnison: ta!11:37
Kinnisonwant review here or by email?11:37
petefothKinnison: whichever is easiest for you11:37
* Kinnison clones11:38
KinnisonArgh, non wordwrapped text11:38
* Kinnison sobs11:38
* Kinnison does it the less pleasant way11:38
* petefoth and Kinnison arfe clearly incompatible on the text-wrapping issue :( 11:39
KinnisonIt makes it *very* hard to review and very hard to understand diffs if the text isn't hard-wrapped11:39
KinnisonThis is why we have line-length limits in code11:39
straycatMutually conflicting versions of eggs are apparently fine, so perhaps that's why they haven't bothered with dependency resolution.11:42
franredstraycat, you mean that different version of python packages can live in the same system?11:43
straycatIf the package is distributed as an egg then apparently yes11:44
radiofreehmm this 3.18 kernel fails to mount the emmc11:50
petefothssam2: there’s already a story in for updating the Trove Reference manual. Ive changed the line ‘need to add building and updating’ to ‘need to add *installing*, building and updating’, which should cover what we discussed11:50
ssam2straycat: the same is true of RubyGems, but you can still have a case where there is no valid dependency graph for a specific version of a Gem. E.g. if it depends on A and B and both require differing versions of C11:50
petefothssam2: not yet psuhed though11:51
ssam2petefoth: OK, I'm planning on sending an email to baserock-dev@ summarising the problems11:51
petefothssam2: great!11:51
ssam2which should pretty much all fall under that story11:51
Kinnisonpetefoth: I remain unconvinced about the last sentence in the Chunk section11:52
Kinnisonpetefoth: As far as Baserock is concerned a chunk's source comes from one or more git repositories11:52
petefothKinnison: yes - that’s one from the last review that I missed. Good spot!11:53
straycatssam2, good point11:55
Kinnisonpetefoth: in definitions, a stratum defines what chunks to include and also where they come from11:56
Kinnisonpetefoth: not sure if that's worth adding11:56
Kinnisonpetefoth: in deployment, you have a "targte"11:57
* Kinnison still dislikes the distributed build text but is having difficulty coming up with replacement text11:59
petefothKinnison: I also missed the comment about ‘tools typically used” ratehr than “tools needed”12:00
Kinnisonpetefoth: I assumed you'd chosen to ignore that one12:00
petefothKinnison: no - it got lost in  the backscroll :)12:01
* Kinnison also isn't sure about the Git description12:01
Kinnisonwe don't use git to control artifacts12:01
KinnisonNot really12:01
Kinnisonpetefoth: I'd suggest a mention that the git cache server is typically found on the Trove, and also a full stop at the end of the sentence.12:02
KinnisonThen again, fullstoppage seems to be inconsistent in the glossary in general, so perhaps that bit is less important12:02
straycatKinnison, which distbuild text?12:05
petefothstraycat: we’re discussing / revieiowing the latest
straycatI'd use the word parallel rather than concurrent.12:06
KinnisonAgain I'm concerned by the text for staging area, but unsure how to clarify it12:07
petefothstraycat: ta!12:07
* Kinnison remains against the lego/duplo analogy12:08
straycatThere's some existing text on distbuild here as well12:08
* pedroalvarez announces:
franredpedroalvarez, \o/12:10
franredcheers :)12:11
Kinnisonpetefoth: otherwise it seems okay12:11
* straycat likes the colours12:12
straycatthey happen to fit the scheme i use for everything else12:12
straycatwell, roughly12:12
pedroalvarezstraycat: I was actually thinking in change some, but keeping the same colours12:12
pedroalvarezs/in change/of changing/ 12:13
abdulssam2: i did some changes in morph file and try to rebuild the image.  it is taking long time to start building that particular morph file package lets say kernel12:15
ssam2abdul: ok, as a starting point could you paste the console output somewhere?12:19
ssam2maybe to the new :)12:19
Kinnisonpetefoth: Let me know if you ever rationalise the raw content enough that I can review it properly :-)12:19
pedroalvarezssam2: how do you use it from the console?12:21
ssam2pedroalvarez: read :)12:23
pedroalvarezaha! it says that it needs a client, but we can use this one:
pedroalvarezradiofree: oh yeah, I have to change the client to show instead of the ip12:28
ssam2abdul: thanks!12:30
ssam2abdul: that is very slow (about 7 minutes before it starts building), I think mostly due to the slow IO on the system you're building on12:31
ssam2there are some inefficient parts of Morph that we know about, so it could definitely be better12:31
ssam2but I don't know how much faster we'd really be able to make it, if /src is on NFS it's always going to be quite slow12:32
perryli'm following the wiki setup instructions and having trouble with one of the commands; 'mkfs.btrfs -L src /dev/sdb' throws a 'no such file or directory' error12:33
pedroalvarezperryl: I think it's because you haven't attached a secondary disk to the vm12:33
ssam2abdul: it's great to have timings, anyway, if I get the chance I might try building over NFS myself and see if there are any obvious improvements to be made12:33
pedroalvarezperryl: what are you using, virtualbox. libvirt, openstack?12:34
pedroalvareza jetson tk1, docker?12:35
perrylpedroalvarez: i think kvm via virt-manager12:35
pedroalvarezperryl: aha, it's easy to add another disk using virt-manager12:36
pedroalvarezperryl: let me know if you need help with that12:36
petefoth111:53 < Kinnison> petefoth: As far as Baserock is concerned a chunk's source comes from one or more git repositories12:36
petefoth1sorry - paste into wrong window :(12:36
*** dutch [] has joined #baserock12:40
abdulssam2, but why it is doing "cp -a /src/cache/gits/git___git_baserock_org_delta_linux /src/tmp/staging/tmp6Pe_qq/", "git checkout 9573986cf27ff44814f3efb1f1e95638406db456", "git cat-file blob" again and again which is consuming more time12:40
abdulssam2, even if i make a small change12:41
*** dutch [] has quit [Client Quit]12:42
ssam2abdul: this is because Morph does a clean build every time, and constructs a clean staging area to do the build in every time12:56
ssam2all of the 'git rev-parse' and 'git cat-file blob' stuff is calculating the build graph, that part is waay too slow and on our list of things to fix12:57
ssam2the 'cp -a' could be replaced if we used some kind of 'union filesystem' to create the build area. That's a bit further in the future though12:58
richard_mawoops, wrong window13:02
petefothKinnison: (and straycat) biff in response to the comments13:06
Kinnisonpetefoth: new format is much easier to read/review \o/13:34
petefothKinnison: I will try to remember to stick with that editor configuration when I’m working on stuff that is likely to be subject to change or review (despit it neign a pain in the arse for most of what I do)13:35
KinnisonPeople tell me off when I whinge about websites having narrow columns, and yet tell me off when I whinge about non wrapped lines in plain text13:36
petefothMost proper modern tools will handle wrapping for you ;)13:37
* petefoth ducks13:37
KinnisonYeah, but then people whinge when my review comments start with "I have reflowed your document to make it easier to review"13:37
KinnisonAlso my nice proper modern tool is explicitly set to not wrap lines13:38
*** tiagogomes [] has quit [Ping timeout: 260 seconds]14:13
*** rdale [] has joined #baserock14:14
*** cosm [] has quit [Ping timeout: 272 seconds]14:23
radiofreepedroalvarez: ok, so found and am testing a 3.18-rc3 based kernel that seems to work pretty well :D14:35
radiofree/dev/mmcblk0p1 / btrfs rw,noatime,ssd,space_cache 0 014:36
pedroalvarezradiofree: great!14:36
*** cosm [] has joined #baserock14:39
*** tiagogomes [] has joined #baserock14:41
radiofreeis it fine for me to keep creating varius kernel branches for jetsons?14:44
rjekWhere might I find information about the process of creating and applying updates/upgrades to a BR system?  I can't spot anything appropriate on the wiki14:44
radiofreethis one is 3.18-rc3 + gpu + cpufreq14:44
radiofreeso we're getting closer to just mainline14:44
Kinnisonradiofree: cool14:45
ssam2rjek: if there's nothing on the wiki that matches what you want, the information probably only resides in people's heads14:47
* rjek gets the trepanning tools out.14:47
ssam2`morph help deploy` has relevant information, probably not in a very useful format14:47
ssam2`morph help ssh-rsync.write` would be useful except that nobody wrote it yet14:47
KinnisonThe Baserock project, at this time, has no end-to-end for creating updates independently of deploying them to a target device14:48
petefothssam2: shoudl i add sh-rsync to the list of write extensions that need documenting?14:48
rjekwhat's the likelyhood of me able able to upgrade my 14.26 VM to 14.40?14:48
ssam2petefoth: yes please!14:48
ssam2rjek: quite high, many of us do that14:48
ssam2rjek: see
* rjek assumes this is not just a case of typing "morph upgrade" in the VM14:49
petefothssam2: Done - see
ssam2rjek: not yet14:49
ssam2the script I linked to shows how to deploy an upgrade with version label TEST14:49
* rjek isn't sure what that means.14:50
ssam2if you do the same but change version label to, say, '14.40', and make sure you run it in a system-branch created with `morph checkout baserock:baserock/definitions 14.40.1`, you should build and upgrade to the 14.40.1 release14:50
ssam2I don't think I can convey more information through IRC, but I'm happy to explain in person if you want14:50
rjekAh, I'm just running a downloaded image from baserock.org14:53
rjekI don't have any branches or anything, I just wanted to upgrade what I already have while keeping the files I'd changed in /etc and such14:54
ssam2currently the only way of upgrading that we have requires you to have used 'morph build' to build the new system14:58
pedroalvarezit should be fast to build it since everything is built in g.b.o14:59
ssam2we seem to generally consider the released images solely as a way of bootstrapping a devel environment from nothing14:59
ssam2so I imagine nobody I am aware of will work on implementing an 'upgrade-from-built-image' mechanism for Baserock15:00
ssam2although I personally would welcome one if someone did create such a thing15:00
*** tiagogomes [] has quit [Quit: Leaving]15:01
*** tiagogomes [] has joined #baserock15:02
straycat is now known as vletrmx2115:09
*** zoli__ [] has quit [Remote host closed the connection]15:10
perrylpedroalvarez: am i right in thinking that to add storage to a vm with virt-manager, i just need to set up a storage pool with a volume inside, or is there more to it?15:18
richard_mawyou can also press something like "browse local" which lets you pick a file that isn't in any pre-defined storage pool15:19
* Kinnison tends to set up an LVM storage pool out of his spare disk and then use that, since those seem to perform better15:20
pedroalvarezperryl: I think that the "browse local" way is the easiest15:21
*** rdale [] has quit [Ping timeout: 245 seconds]15:34
paulsher1oodpedroalvarez: i've tried your patch but it doesn't seem to be working for me15:40
pedroalvarezpaulsher1ood: same error?15:41
paulsher1oodyup... i'm wondering if i've somehow confused morph altogether15:41
paulsher1oodwhat does morph --version return these days?15:42
pedroalvarezthe sha115:42
paulsher1oodah, it's me15:42
paulsher1oodbecause i've booted into a new system, i'm getting the system morph :)15:42
*** moss_maurice [] has joined #baserock15:43
vletrmx21hello moss_maurice15:43
moss_mauricehi vletrmx2115:43
pedroalvarezpaulsher1ood: let me know if it works 15:44
moss_mauriceI'm trying to set up a baserock VM, but the fallocate command from the wiki gives me "fallocate failed: Operation not supported", is there a nicer way of allocating space then with dd?15:48
paulsher1oodpedroalvarez: it does ! +115:48
Kinnisonmoss_maurice: You could try truncate(1) instead of fallocate(1)15:49
Kinnisonalthough I don't think that'll allocate the blocks, just set the size of the file15:49
moss_mauriceok, I'll give it a try.15:50
pedroalvarezpaulsher1ood: great!15:50
* pedroalvarez merges15:50
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock15:56
perryli'm noticing issues with rebooting my baserock vm in accordance with these instructions:
persiaperryl: What sort of issues?16:04
perryli'm not sure if it's mkfs.btrfs or modifying /etc/fstab but upon rebooting i get a timeout message waiting for device dev-disk-by\x2dlabel-src.device, and it boots to emergency mode16:04
perryli'm guessing with label-src it's the line added to /etc/fstab?16:05
pedroalvarezperryl: yup16:05
Kinnisonperryl: when you did the mkfs, did you label the disc?16:05
perrylKinnison: i inputted the commands exactly as they're written on the wiki; is label defined as -L?16:07
KinnisonI believe so16:07
perrylin that case yes16:07
radiofreeperryl: we need to patch the defaults (wherever they are used..) to use nofail16:07
*** moss_maurice [] has quit [Quit: Leaving]16:09
perrylit might help if after -L i include 'src'...thanks, Kinnison!16:09
*** rdale [] has joined #baserock16:11
*** moss_maurice [] has joined #baserock16:13
petefoth1perryl: please update the nof-frills wiki page if what's writen there doesn't work for you16:54
perrylpetefoth1: it was me mistyping, i believe no-frills is currently fine :)16:55
petefoth1perryl: :) That's all right then!16:56
Kinnisonfranred: regarding dmidecode -- eww CVS16:57
franredKinnison, ummm, I found a git repo:
franredfrom --> "and is now maintained by Anton Arapov"17:01
Kinnisonthat seems to lack recent commits to the CVS17:01
franredummm yep :/17:02
franredthe CVS repository is from:
franredwhich is the one which they post in their webpage, so I think I will stick with CVS for the moment17:03
radiofreehi jjardon,what was the GDK env to always use opengl?17:05
jjardonradiofree: GDK_ALWAYS_USE_GL17:07
jjardonKinnison: about gcc: thanks! I only though it could be useful for some dev in the future, but I guess a tarball is fine for now17:09
*** jonathanmaw [] has quit [Quit: Leaving]17:13
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]17:14
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock17:24
*** franred [] has quit [Quit: Leaving]17:25
*** moss_maurice [] has quit [Quit: Leaving]17:52
*** mariaderidder [] has quit [Quit: Ex-Chat]17:59
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]18:03
*** Guest39703 [] has quit [Quit: Leaving]18:03
*** tiagogomes [] has quit [Quit: Leaving]18:21
vletrmx21 is now known as straycat18:38
*** ssam2 [] has quit [Quit: Leaving]19:05
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock19:13
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]19:25
*** genii [~quassel@ubuntu/member/genii] has joined #baserock19:52
*** rdale [] has quit [Remote host closed the connection]20:33
*** rdale [] has joined #baserock20:37
*** zoli_ [] has joined #baserock20:43
*** zoli_ [] has quit [Changing host]20:43
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock20:44
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]20:47
*** abdul [~abdul@] has quit [Remote host closed the connection]21:06
robtaylorpaulsher1ood: you may be interested to know that cosm has figured out how to upgrade the jetson to 4g :)22:24
cosmyeah, actually it's been running a memory tester for about 24hrs now22:28
*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]22:28
robtaylorcosm: nice!22:29

Generated by 2.15.3 by Marius Gedminas - find it at!