#jogamp @ irc.freenode.net - 20130604 05:05:44 (UTC)


20130604 05:05:44 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130603050530.html
20130604 05:05:44 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130604050544.html
20130604 08:16:30 <sgothel> http://jogamp.org/git/?p=jogamp-scripting.git;a=blob;f=server/setup/zfs_linux_bringup/README.txt;h=2eac52fd9b230686633a145b4df6ce6be0543efa;hb=HEAD <- setting up .. (took quite some time to figure out whats working and whats not .. well)
20130604 08:16:57 <sgothel> soon this setup will be easier, when grub works well w/ zfs
20130604 08:37:19 * [Mike] (~Mike]@anon) Quit ()
20130604 09:53:05 * petrs (~petrs@anon) Quit (Read error: Connection reset by peer)
20130604 09:53:51 * petrs (~petrs@anon) has joined #jogamp
20130604 11:36:45 * jk4 (~jk4@anon) Quit (Ping timeout: 245 seconds)
20130604 11:37:02 * jk4 (~jk4@anon) has joined #jogamp
20130604 12:04:38 <sgothel> populating .. jogamp.org -> jogamp_b (new server)
20130604 12:05:30 * rmk0 holds breath
20130604 12:06:25 <sgothel> http://jogamp.org/git/?p=jogamp-scripting.git;a=blob;f=server/setup/zfs_linux_bringup/README.txt;hb=HEAD .. updated
20130604 12:06:37 * rmk0 eyes it
20130604 12:07:47 <sgothel> well .. after reading through the night(s) .. how horrible the state of zfs on linux might be .. ( https://github.com/zfsonlinux/zfs/issues?labels=Bug&milestone=22&page=1&state=open ) .. I jumped to the believe .. and .. well, just did it :)
20130604 12:08:22 <sgothel> so .. backup .. is the magic word for now .. :)
20130604 12:09:23 <rmk0> ugh
20130604 12:09:46 <rmk0> that is massively frustrating
20130604 12:09:48 <sgothel> before I switch completely (DNS) .. I might not be able to test all services, since they rely on DNS
20130604 12:09:52 <sgothel> .. yes :)
20130604 12:09:53 <rmk0> can't get over the stupidity of the situation
20130604 12:10:07 <rmk0> oracle won't relicense zfs, so it'll never be included with the kernel
20130604 12:10:21 <rmk0> and instead, they spend thousands of man hours developing a completely awful non-alternative
20130604 12:10:41 <sgothel> at least .. it seems that illumos, freebsd and linux team is cooperating .. hopefully well
20130604 12:11:03 <sgothel> when I was testing 3 grub/zfs repos .. it didn't look that well .. i.e. patches patches patches ..
20130604 12:11:26 <rmk0> bleh
20130604 12:11:40 <sgothel> hence I gave up and use zfs-initramfs .. with manual hacks to have it even boot. good think: learned new stuff :)
20130604 12:12:13 <sgothel> freeze etc is ok, as long no data is gone .. the latter class of bugs seem to be limited :)
20130604 12:12:32 <xranby> how do btrfs stand in comparation to zfs now days?
20130604 12:12:42 <sgothel> lol
20130604 12:13:05 <rmk0> it's still totally unusable
20130604 12:13:18 <sgothel> at least zfs has many years of development .. core etc, and the ports stuff (illumous, freebsd, linux) .. 2-3 ?
20130604 12:13:43 <sgothel> so .. maturity wise, it's the best .. I only would wish, xfs would support those checksums .. :(
20130604 12:14:18 <rmk0> i'd take the checksums and volume/pool management
20130604 12:14:35 <sgothel> sha256 + raid-ish-something :)
20130604 12:14:45 <rmk0> is really nice to be able to create and delete filesystems and snapshots as if they were nothing
20130604 12:14:53 <sgothel> yup
20130604 12:14:57 <sgothel> like git on fs :)
20130604 12:15:01 <rmk0> pretty much
20130604 12:15:15 <rmk0> my backup server here is producing nightly filesystem snapshots
20130604 12:15:16 <sgothel> btw .. does git checksum ?
20130604 12:15:34 <rmk0> sort of
20130604 12:15:40 <sgothel> i.e. can validate if bits got toggled ?
20130604 12:16:08 <rmk0> when you commit something, the contents of the commit are hashed with sha-1 and then that "blob" is stored using the sha-1 as a name
20130604 12:16:18 <rmk0> but i'm not sure when/if it validates old blobs
20130604 12:24:53 * xranby hope all the mirroring raid etc work in practise.. that is will the system still boot when a) a drive gets unplugged/replugged while power on, b) a drive gets removed while power offf
20130604 12:28:10 <xranby> http://www.adrc.com/raid_failure_types.html
20130604 12:31:17 <xranby> sgothel: running a backup server looks still like the best solution
20130604 12:34:25 <sgothel> yup. as long server has ECC ram, the sha256+atom_write+redundancy is considered (proven) safe - at least the theory :)
20130604 12:34:46 <sgothel> I actually unlugged my local test machine and did a few cycles .. works
20130604 12:34:53 <xranby> cool
20130604 12:35:00 <sgothel> the only weak link right now is the USB drive to boot
20130604 12:35:26 <xranby> can the USB drive boot and start the system using any of the two disks?
20130604 12:35:29 <sgothel> but since no write operation is performed on it .. (normally) it should be ok
20130604 12:35:41 <sgothel> the 2 disks are one zfs wholedisk
20130604 12:35:46 <sgothel> mirror setup
20130604 12:35:55 <sgothel> (raid1)
20130604 12:36:19 <sgothel> zfs create a GPT partition w/ fake MBR
20130604 12:37:15 <sgothel> so any 1 can fail
20130604 12:37:47 <sgothel> at home I will have raidz1 (2 + 1), allowing 1/3 to fail
20130604 12:38:49 <sgothel> backups will be run regularly .. yup
20130604 14:33:34 <sgothel> http://www.theregister.co.uk/2013/06/04/chinas_internet_maintenance_day_shutters_sites/ <- China's Internet Maintenance Day .. maybe a good day to shut off jogamp for .. err .. maintenance ? :)
20130604 17:07:27 <sgothel> 1st rsync finished!
20130604 17:08:16 <sgothel> until tomorrow evening ( I will ping you ) .. please do not write to jogamp, git, bugzilla, email .. - is that ok ?
20130604 17:08:38 <rmk0> right
20130604 17:09:06 <rmk0> .. might be worth putting that in the irc topic
20130604 17:11:14 <sgothel> I start w/ our own services .. http/git/jenkins at last tomorrow
20130604 17:11:38 <sgothel> our own (users/email/...) .. or maybe I get some sleep and do it fresh tomorrow :)
20130604 17:29:18 <sgothel> @Xerxes: Thank you for helping sorting out (managing) bugreports, mailinglist .. lately - good job. I just dared to look at the forum emails and bug reports again :)
20130604 17:29:34 <sgothel> .. just in case I finish the server task soon :)
20130604 17:45:13 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=742
20130604 17:45:38 * hharrison (~chatzilla@anon) has joined #jogamp
20130604 17:46:56 <hharrison> sgothel: just getting to quaternion patches, had a fedora-upgrade catastrophe
20130604 17:47:32 <sgothel> thank you that you do that .. !
20130604 17:47:38 <sgothel> which kind of ?
20130604 17:47:42 <sgothel> what FS ?
20130604 17:47:57 <sgothel> since .. we can only learn .. you know .. me doing https://jogamp.org/bugzilla/show_bug.cgi?id=742
20130604 17:49:17 <sgothel> looks like it's the time of the year .. maintaining servers .. :)
20130604 17:50:01 <sgothel> hope it's ok that I have included the fw settings in git :)
20130604 17:50:28 <sgothel> if anybody finds a bug in anything (setup) .. pls fix .. reply
20130604 17:51:19 <sgothel> Harvey .. are you in .. Siggraph ? Demo ? Slides ? I guess in 1-2 weeks I will start bugging again for slides :)
20130604 17:56:02 <hharrison> for siggraph: slides + demo for sure
20130604 17:56:36 <hharrison> I'm possibly committed to attend a wedding the SUnday immediately before siggraph
20130604 17:56:59 * xranby1 (~xranby@anon) has joined #jogamp
20130604 17:57:02 <hharrison> Trying to nail down the date on that, otherwise I will be there for the siggraph BOF
20130604 17:57:24 <hharrison> WRT filesystems, I'm a ext4 or xfs guy
20130604 17:57:43 <hharrison> And then I track anything I actually care about in git
20130604 17:59:14 <sgothel> we just mentioned it .. git does not always check sha's when doing i/o right ?
20130604 17:59:19 <sgothel> (zfs does ..)
20130604 17:59:33 * xranby (~xranby@anon) Quit (Ping timeout: 264 seconds)
20130604 17:59:35 <sgothel> so .. is there a git scrubbing command ?
20130604 17:59:47 <sgothel> i.e. validating whether the git repo is still alive and kicking ?
20130604 18:00:00 <sgothel> w/o silent damages ..
20130604 18:01:08 <hharrison> git fsk
20130604 18:01:11 <hharrison> git fsck
20130604 18:01:20 <hharrison> checks the graph connectivity
20130604 18:01:24 <hharrison> git fsck --full
20130604 18:01:31 <hharrison> checks the full blob consistency
20130604 18:01:40 <sgothel> ah .. thx
20130604 18:02:06 <hharrison> But I tend to then do occassional repacking, write a .keep file for that pack and archive it to long term storage separaetly
20130604 18:02:19 <hharrison> Say, every 3-6 months or so
20130604 18:02:42 <sgothel> any magic script doing this recursively from a top git dir ?
20130604 18:03:32 <hharrison> Not really, I have mine setup so that there is a public 'shared' git repo that I pack/archive carefully
20130604 18:03:50 <hharrison> Then the satellite repos just reference that one to pull objects from
20130604 18:04:13 <sgothel> oh .. ok, will add it to crontab then .. thx
20130604 18:04:51 <hharrison> I will add some details to the referenced bugzilla above over the next few days....in case you find anything useful from my setup
20130604 18:05:07 <sgothel> ofc .. thx
20130604 18:05:23 <hharrison> The nice part is the shared repo is marked totally readonly
20130604 18:08:01 * jk4 (~jk4@anon) Quit (Ping timeout: 252 seconds)
20130604 18:12:08 * jk4 (~jk4@anon) has joined #jogamp
20130604 18:32:59 -mquin- [Global Notice] As you've probably noticed we're experiencing a bit of network disruption. Please bear with us while we check the tubes for kittens.
20130604 18:34:35 * odin_ (~Odin@anon) Quit (Excess Flood)
20130604 18:37:40 * jk4 (~jk4@anon) Quit (Ping timeout: 245 seconds)
20130604 18:38:04 <hharrison> Things look OK in those quaternion patches, would prefer if petr cleaned up his commiter name a bit
20130604 18:38:29 <sgothel> I agree .. just ask him, I guess he is motivated.
20130604 18:39:01 <sgothel> I won't touch server things until I get up tomorrow (in ~12h) .. so you still can write :)
20130604 18:39:20 * jk4 (~jk4@anon) has joined #jogamp
20130604 18:40:35 * hharrison (~chatzilla@anon) Quit (Read error: Connection reset by peer)
20130604 18:40:37 * hharrison (~chatzilla@anon) has joined #jogamp
20130604 18:40:37 * hharrison (~chatzilla@anon) Quit (Client Quit)
20130604 18:40:37 * hharrison (~chatzilla@anon) has joined #jogamp
20130604 18:41:07 <hharrison> I need to have a look at the new mult() function he introduces in his final patch, but it looks at first read to be OK
20130604 18:41:21 <hharrison> Good enough that we can do any fixups later if needed
20130604 18:45:44 <hharrison> I commented on github in his repo, never sure then best way to do code review without a mailing list
20130604 18:45:54 <hharrison> \old man mode
20130604 18:46:29 <sgothel> additional pm maybe / ah .. you do a bugzilla note I guess, I always miss github notifications :)
20130604 18:47:00 <hharrison> Sure, I'll copy the comments into the bugzilla, hope he doesn'y mind getting it twice.
20130604 18:50:35 <hharrison> done
20130604 18:59:06 <sgothel> thx ..
20130604 19:04:04 * hharrison_ (~chatzilla@anon) has joined #jogamp
20130604 19:04:35 * hharrison_ (~chatzilla@anon) Quit (Client Quit)
20130604 19:05:25 * hharrison_ (~chatzilla@anon) has joined #jogamp
20130604 19:05:36 * hharrison_ (~chatzilla@anon) Quit (Client Quit)
20130604 19:05:39 * hharrison (~chatzilla@anon) Quit (Ping timeout: 245 seconds)
20130604 20:40:25 * hharrison (~chatzilla@anon) has joined #jogamp
20130604 21:38:01 * orthoplex64 (~orthoplex@anon) has joined #jogamp
20130604 23:55:39 * [Mike] (~Mike]@anon) has joined #jogamp
20130604 23:56:37 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90 [Firefox 22.0/20130528181031])
20130605 01:43:01 * orthoplex64 (~orthoplex@anon) Quit (Ping timeout: 246 seconds)
20130605 02:39:51 * orthoplex64 (~orthoplex@anon) has joined #jogamp
20130605 02:46:59 * orthoplex64 (~orthoplex@anon) Quit (Ping timeout: 260 seconds)
20130605 03:05:10 * [Mike] (~Mike]@anon) Quit ()
20130605 03:08:26 * orthoplex64 (~orthoplex@anon) has joined #jogamp
20130605 03:19:58 * orthoplex64 (~orthoplex@anon) Quit (Ping timeout: 246 seconds)
20130605 03:39:03 * jk4_ (~jk4@anon) has joined #jogamp
20130605 03:41:40 * jk4 (~jk4@anon) Quit (Remote host closed the connection)
20130605 04:16:12 * orthoplex64 (~orthoplex@anon) has joined #jogamp
20130605 04:22:37 * orthoplex64 (~orthoplex@anon) Quit (Ping timeout: 246 seconds)
20130605 05:05:57 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130605050557.html