[01:20] <grillo> Hi.  I'm setting up new drives, headed for a RAID array for my new UbuServer.  They're set up as GPT.  For now, mobo's non-EFI -- that'll eventually change.  For now, though, I need a "BIOS boot partition" to make Grub2 happy.  IIUC, that needs to be at the "start" of the drive(s).  Question -- at which sector should it start?  sector=1?, the UNALIGNED start of the largest free block (sgdisk -f == 34), or the ALIGNED start of the largest free block
[01:20] <grillo> (sgdisk -F == 2048)?
[06:36] <lordievader[m]> Good morning
[07:34] <TafThorne> morning
[08:44] <yossarianuk> ‎ hi -  I am setting up a replicated postgresql server, using pacemaker for HA - its for a Zaabix server, what advantage would also using DRDB give me ?
[08:45] <yossarianuk> many guides I have seen suggest using it - just wondering what advantage it would give?
[08:46] <maswan> Hm. That seems like layering two node replication steps on top of eachother
[08:47] <maswan> I don't see what you'd gain, other than more complexity and a potential fault source
[08:56] <yossarianuk> ‎maswan: thanks for the response
[08:56] <fishcooker> yossarianuk: how about failover scene?
[08:57] <yossarianuk> ‎maswan: That was my thought also, was wondering is perhaps there was better performance writing to a network shared block device over replicating normally..
[08:59] <maswan> Should be worse actually, since DRBD requires each block write to hit disk on both places before returning from a write, rather than postgres just managing transaction syncs
[09:00] <maswan> But I haven't benchmarked it head-to-head, so that's just theory from my side
[09:01] <maswan> We run a fairly big postgres with repmgr and manual failover (we only have two servers there, so we can't rely on getting quorum)
[09:01] <yossarianuk> ‎fishcooker: I am using this as a template for HA -> http://wiki.clusterlabs.org/wiki/PgSQL_Replicated_Cluster
[09:06] <yossarianuk> ‎maswan‎: thanks - it was guides like this that made me consider DRDB -> https://wiki.postgresql.org/images/0/07/Ha_postgres.pdf
[09:06] <yossarianuk> Its for a zabbix server setup so pretty sure pacemaker should be fine in that case.
[09:09] <maswan> postgres 9.1 is pretty old in terms of replication support
[09:10] <maswan> streaming replication went into 9.3 I think? and it's much better in 9.5+
[09:14] <fishcooker> noted, yossarianuk
[09:22] <yossarianuk> i think we plan to use 9.2, would it be better to bump the version in that case?
[09:23] <maswan> wow, that's old. xenial has 9.5 and trusty on 9.3
[09:25] <yossarianuk> I think its being compiled from source (for some reason I am unaware of) - i'll speak to them and see why and encourage a newer version..
[09:25] <yossarianuk> thanks for the advice... it a fairly new area for me, I usually use Mysql.
[15:14] <aoam> hello, i have a little question, is it possible to run systemd system in lxd container on host that has no systemd ? (because of im having troubles with that, servicies inside doesnt work)
[17:29] <nacc> powersj: LP: #1583126, src:mysql-5.6 is in universe in trusty, so we can unsub ubuntu-server?
[17:58] <smoser> nacc, https://bugs.launchpad.net/ubuntu/+source/open-iscsi/+bug/1666573
[23:22] <dpb1> teward: took a look at the release note blurb for nginx, looks great. :)
[23:27] <teward> dpb1: glad to hear it :)
[23:27] <teward> anything that needed changed, or is it good as is?
[23:28] <nacc> teward: as is, is great