| 01:25.23 | *** join/#edev uwe_ (~uwe_@dslb-188-105-024-052.pools.arcor-ip.net) |
| 03:05.15 | *** join/#edev _Lucretia__ (~munkee@5ac2a2af.bb.sky.com) |
| 04:39.32 | *** join/#edev TheSeven (~quassel@rockbox/developer/TheSeven) |
| 05:06.44 | *** join/#edev mbuf (user@nat/redhat/x-sjlpttdmlzwmasud) |
| 06:49.00 | *** join/#edev clio (~andrej@85.159.109.222) |
| 09:30.12 | *** join/#edev keesj (~keesj@dellpc132.few.vu.nl) |
| 09:39.00 | *** join/#edev jackmitchell (~Thunderbi@195.171.99.130) |
| 09:39.43 | *** join/#edev florian (~fuchs@82.113.121.241) |
| 09:39.43 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 10:06.26 | *** join/#edev calculu5 (~calculus@gentoo/user/calculus) |
| 10:18.31 | *** join/#edev florian (~fuchs@89.204.139.59) |
| 10:18.31 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 10:55.59 | *** join/#edev sjhill (~sjhill@home.bethel-hill.org) |
| 10:55.59 | *** mode/#edev [+o sjhill] by ChanServ |
| 10:56.35 | *** join/#edev soafee-chan (~spaec@unaffiliated/traumapony) |
| 11:14.31 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 11:54.52 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 12:05.13 | *** join/#edev Shinda (~Loup@91.183.199.211) |
| 12:05.13 | *** join/#edev Shinda (~Loup@unaffiliated/toroop) |
| 12:05.41 | Shinda | hello |
| 12:24.51 | *** join/#edev gustavoz (~gustavoz@host110.190-225-90.telecom.net.ar) |
| 12:25.55 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 12:25.55 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 12:26.04 | *** join/#edev g1powermac_ (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 12:53.41 | *** join/#edev _Lucretia_ (~munkee@pdpc/supporter/active/lucretia) |
| 12:53.53 | *** join/#edev florian (~fuchs@89.204.154.163) |
| 12:53.53 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 12:54.39 | *** join/#edev jawilson (jawilson@2600:3c03::f03c:91ff:fe93:fbce) |
| 13:50.35 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 14:00.56 | *** join/#edev Shinda (~Loup@91.183.199.211) |
| 14:00.57 | *** join/#edev Shinda (~Loup@unaffiliated/toroop) |
| 14:04.03 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 14:04.03 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:53.51 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 15:55.19 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
| 16:04.13 | *** join/#edev prpplague (~danders@192.91.66.186) |
| 16:16.18 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
| 16:49.12 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-50def8-126.dhcp.inet.fi) |
| 17:02.20 | *** join/#edev is0-mick (~is0-mick@40.27.125.91.dyn.plus.net) |
| 17:24.27 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 17:25.17 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 18:01.02 | *** join/#edev Belliash (belliash@asiotec/constitutive/belliash) |
| 18:06.12 | *** join/#edev megha (~fire@unaffiliated/security) |
| 18:49.40 | *** join/#edev bzb (~bzb@69-196-189-45.dsl.teksavvy.com) |
| 19:04.57 | *** join/#edev aarias (~aarias@48.Red-80-59-144.staticIP.rima-tde.net) |
| 19:05.54 | *** join/#edev spidercaleb (~caleb@tx-71-1-45-3.sta.embarqhsd.net) |
| 19:06.05 | spidercaleb | Does anyone here know much about aufs? |
| 19:19.35 | wmat | spidercaleb: I thought aufs was essentially dead |
| 19:20.02 | spidercaleb | Anything better? |
| 19:22.25 | wmat | there was UnionFS, if I remember correctly |
| 19:22.48 | wmat | http://www.filesystems.org/project-unionfs.html |
| 19:24.14 | wmat | there's lots more info here: http://elinux.org/File_Systems |
| 19:27.53 | spidercaleb | From what I've read, aufs is better (I don't remember why). |
| 19:28.00 | spidercaleb | But I'll check it out |
| 19:28.18 | spidercaleb | pun intended :P |
| 19:28.31 | gustavoz | wmat: unionfs (non-fuse) seems pretty much dead too |
| 19:28.46 | gustavoz | there's unionfs-fuse (slow) or overlayfs |
| 19:29.21 | spidercaleb | ...that too |
| 19:29.32 | wmat | yes, those are all listed on the elinux.org page, aren't they |
| 19:29.37 | gustavoz | overlayfs isn't |
| 19:30.19 | wmat | mini_fo is similar |
| 19:30.34 | gustavoz | is it maintained other than by openwrt people? |
| 19:30.48 | wmat | there must be a good reason why all of these died |
| 19:31.07 | gustavoz | actually openwrt switched to overlayfs :) |
| 19:31.28 | gustavoz | i switched to overlayfs from unionfs since they started lagging in updates until i've seen no more |
| 19:31.45 | wmat | interesting |
| 19:32.11 | wmat | was the feature set further along than overlayfs at the time you switched? |
| 19:33.14 | gustavoz | my use case is pretty simple actually, some RO flash that has a default config and an overlay RW filesystem (jffs2 or other) that supercedes the defaults, usually for /etc |
| 19:33.17 | kergoth | i gave up on keeping up with the latest news in union filesystems, seems like it's always changing |
| 19:33.35 | gustavoz | so when a config reset is wanted just zap the RW filesystem |
| 19:34.12 | kergoth | that's a common case, indeed |
| 19:34.23 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 19:34.34 | gustavoz | one could presumably handle multiple config versions with more than one JFFS rotating |
| 19:36.01 | spidercaleb | like preset configurations for certian roles? |
| 19:36.21 | gustavoz | possibly, or fallback know-working configurations |
| 19:36.25 | gustavoz | *known even |
| 19:37.23 | spidercaleb | Like xfs? |
| 19:37.29 | spidercaleb | Can overlayfs do that? |
| 19:38.41 | spidercaleb | It doesn't really matter to me; I won't be using it like that. |
| 19:38.54 | spidercaleb | My question is: if I change a file on the ro branch and it is in use on the union mount, what happens? |
| 19:41.52 | gustavoz | if that's the only file in the different branches it'll just work |
| 19:42.35 | gustavoz | otherwise it must follow the rules |
| 19:43.55 | spidercaleb | Latest created/modified file takes precedence? |
| 19:44.30 | gustavoz | if you've got conflicting files that would depend on the branch order you tell the union filesystem |
| 19:45.46 | gustavoz | in overlayfs the upper filesystems override the lower ones, say if the lowest one is the RO filesystem (typical use) and an upper has a newer file then the upper wins, or if it's deleted in the upper (uses a special filename for that) the file is gone |
| 19:45.50 | gustavoz | well "gone" |
| 19:46.28 | spidercaleb | So in dirs=1:2=ro , a file in 1 would overrule one in 2? |
| 19:46.49 | spidercaleb | and deleting that file would make it ignore the one in 2/ ? |
| 19:47.56 | gustavoz | see http://git.kernel.org/?p=linux/kernel/git/mszeredi/vfs.git;a=commitdiff;h=38c4f55850b118899c10a2811cd436b2d051303a for the details |
| 19:49.38 | gustavoz | from there it seems NFS might be problematic |
| 19:50.20 | spidercaleb | How so? |
| 19:50.48 | gustavoz | when using NFS as the read-write component it's a problem |
| 19:51.08 | gustavoz | IIRC it uses extended attributes to denote for example deleted files |
| 20:34.12 | spidercaleb | Appearantly overlayfs is having problems with linux 3.7m which is what i'm using for my project |
| 20:35.22 | gustavoz | for aufs you could just convert the colons into - or something else for a quick workaround |
| 20:35.47 | gustavoz | i'm using overlayfs with 3.4 & 3.6 for the moment so i can't say much about 3.7 |
| 20:36.31 | spidercaleb | That's what I'm doing. I didn't think it was worth mentioning ,though |
| 20:36.37 | spidercaleb | Thanks all |
| 20:40.16 | spidercaleb | I actually already have that. I have to take BOOTIF=01-ab-cd-ef-12-34-56 from pxelinux and convert it into ab:cd:ef:12:34:56 |
| 20:40.27 | spidercaleb | I just have to export it |
| 20:40.40 | *** join/#edev LeelooMinai (~leeloo@198-84-185-212.cpe.teksavvy.com) |
| 20:46.34 | gustavoz | that's a quick sed |
| 20:48.15 | *** join/#edev LeelooMinai (~leeloo@198-84-185-212.cpe.teksavvy.com) |
| 20:48.24 | spidercaleb | I'm saying I get xx:yy:zz from xx-yy-zz in the first place |
| 20:48.50 | spidercaleb | So I'm recycling the first string rather than undoing the substitution I did before |
| 20:49.12 | spidercaleb | AND IT ALL WORKS!!! WOOHOO!!! |
| 20:50.16 | spidercaleb | gustavoz: If I delete a file, it disappears. |
| 20:50.34 | spidercaleb | I'm thinking it falls back to special files if it can't set the extended attribute |
| 20:50.54 | gustavoz | spidercaleb: the limitation was for overlayfs, probably not aufs |
| 20:51.27 | spidercaleb | Oh. |
| 20:51.38 | gustavoz | spidercaleb: unionfs uses shadow files, for example deleted hosts would be .dht.del.hosts or some crap like that |
| 20:51.57 | gustavoz | spidercaleb: aufs probably does something similar, LSing around you can probably see the trick |
| 20:52.22 | gustavoz | just look into the RW dir with ls -la to see hidden gems |
| 20:52.43 | spidercaleb | gustavoz: I saw them before which is why I said it might fall back |
| 20:53.19 | gustavoz | spidercaleb: usually they just use one method, IIRC unionfs didn't use extended attributes at all |
| 20:53.45 | gustavoz | overlayfs probably does in order to be simpler and get on kernel upstream good graces to eventually be upstreamed |
| 20:54.19 | spidercaleb | I can see why |
| 20:54.26 | gustavoz | using the filename trick isn't 100% foolproof - you could get filename clashes with bad results |
| 20:54.44 | spidercaleb | The whole shadow file thing is just an ugly hack in the first place |
| 20:54.46 | gustavoz | very fringe, but it could happen |
| 20:54.52 | spidercaleb | /srv/nfs/nodes/08-00-27-e6-06-88/etc# ls -a |
| 20:54.52 | spidercaleb | . .. random-seed .wh.os-release |
| 20:56.01 | spidercaleb | Anyways, thanks again |
| 20:56.39 | spidercaleb | is really looking forward to getting GPU drivers to work now |
| 20:57.05 | gustavoz | you're welcome |
| 21:32.26 | iospace | you know, after doing embedded devel, you sort of get an appreciation for the vast number of pins on chips |
| 21:35.08 | gustavoz | someone threw an attiny at you? :) |
| 21:38.28 | iospace | gustavoz: no, i work with i7s :P |
| 23:02.58 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 23:07.41 | *** join/#edev contempt (contempt@unaffiliated/contempt) |
| 23:17.37 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 23:17.38 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 23:27.55 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |