[12:42] <erle-> is the ZFS in 16.04 a port of  Solaris code or a different implementation?
[12:50] <rtg> erle-, https://github.com/zfsonlinux/zfs.git
[12:51] <erle-> thanks
[12:51] <erle-> that's exactly what I was looking for
[15:48] <rharper> xenial s390x, loading bcache module, I see no /sys/class/block/bcache dir ... is that expected? 
[16:11] <rharper> indeed, I was looking for /sys/fs/bcache; 
[17:31] <pkern> Hi. I seem to have a hard time to get pstore to persist a panic (both with efi and esrt backends). Does anyone happen to know how to turn it on properly? I set "pstore.backend=efi/esrt printk.always_kmsg_dump=Y" to test, but then I guess the latter might not work because CONFIG_PSTORE_CONSOLE isn't set...
[18:14] <stgraber> cking: hey, so zfsutils-linux pulls in zfs-zed right now which then pulls in an e-mail server, that doesn't seem ideal. Could we have zfs-zed be demoted to suggests?
[18:14] <cking> stgraber, yep, sounds like a good idea
[18:14] <cking> i'll do that right now
[18:14] <stgraber> cking: thanks
[18:16] <stgraber> cking: zfs supports s390x now right?
[18:16] <cking> stgraber, s390x, aarch64, powerpc64 and amd64
[18:16] <stgraber> cking: hmm, maybe I'm a bit behind on my ubuntu kernels then, not seeing a zfs.ko on s390x
[18:17] <cking> stgraber, yep, wait for next kernel
[18:17] <stgraber> ok, I see it in -16 but I'm on -15 for some reason
[18:17] <stgraber> lets reboot that box
[19:01] <infinity> stgraber: As of -17 (or -18?), it'll be all 64-bit kernels.
[19:01] <infinity> stgraber: But yeah, I think it's on in -16 on s390x.
[19:02] <stgraber> yeah, I just tested it on s390x with -16 and it works great
[19:03] <infinity> stgraber: Shame about the lack of 32-bit kernel support for people like you who are (I assume) building userspace on top. :/
[19:04] <infinity> stgraber: I guess you're doing some if zfs; elif btrfs; elif overlay; else FUCK; fi thing? :P
[19:04] <stgraber> yeah, we tested it on i386 just for fun a while back, had that machine kernel panic a dozen times in a day :)
[19:05] <stgraber> nah, lxd by default does plain old dir, you have to opt into anything fancier or run "lxd init" which does the setup for you based on a few questions
[19:06] <stgraber> but yeah, we've had to push back on pressure to always setup zfs due to not having it on all arches which means all the docs and stuff have to cover all the storage backends (zfs, btrfs, lvm+ext4, lvm+xfs and plain-old-slow-dir)
[19:07] <infinity> stgraber: Oh, good, lvm too.  You haven't forgotten people like me who live comfortably in the past.
[19:08] <stgraber> yeah, we do lvm with thin provisioning but it kinda sucks when you try to migrate things between machines, doesn't let us resize the container up and down while it's running and is kinda slow in general, but yeah, we have it