10:25:47 <Woodstock> tsoome: is there a reason you stop at 4k sector size for pcfs? is it impossible or too hard to implement 8k, 16k, etc.?
11:03:32 <tsoome> TBH, I was just using list of values from microsoft fatgen103 document, as long as the value is power of 2 -- other than it, the size value does not really change the code itself.
11:05:40 <Woodstock> i see
11:06:36 <tsoome> basically any value above 4k is non-standard, dosfstools mkfs is allowing values up to 32k
11:06:45 <Woodstock> i see
11:07:21 <Woodstock> i was just wondering because in the past i used virtio-blk devices in some VMs that had a 8k sector size, although I don't recall whether that was 8k logical or 8k native
11:08:33 <tsoome> yep, I was also thinking about those 8k native flash memory implementations etc. 
11:10:36 <tsoome> from practical point, the fs itself is depending on 512B structures for boot sector and fsinfo (and their backups), so with 32k sector, you will have number of unused bytes:D
11:12:08 <Woodstock> i'm not losing sleep about wasting 32256 bytes :)
11:12:27 <tsoome> ofc the current loader dosfs does not really check for those backups (nor alternate fat). 
11:12:36 <tsoome> :)
11:14:09 <tsoome> ok. need to fix pcfs now, then I can verify my installboot patch...
11:14:46 <tsoome> then there are mkfs and fsck...
11:17:57 <Woodstock> have fun :)
12:20:08 <tsoome> um, did we fix reboot with iscsi initiators?
13:07:31 <tsoome> nop, we did not.
14:29:45 <jbk> no
14:29:49 <jbk> it's still not great
18:28:44 <richlowe> it seemed like everyone decided what the fix was, and then nobody fixed it
18:29:06 <richlowe> probably because the people who figured out what to do don't have pools on iscsi to check it out?
18:43:45 <jbk> i mean, you could use another illumos instance to serve up an iscsi volume
18:43:59 <jbk> but our iscsi target code is pretty terrible tbh
18:44:28 <jbk> (none of our customers use it, so we've not really had a reason to dedicate time to it)
18:45:21 <jbk> (if you want to see some fireworks, try using PGR on our target code :P)
19:10:23 <tsoome> jbk yea, its the opposite of great:)
19:41:10 <richlowe> "none of our customers use it" sure sounds a bit chicken and egg
19:41:25 <richlowe> "What we sell is crap, because nobody uses it"
19:41:27 <richlowe> I mean, yeah
19:54:42 <jbk> for our stuff it doesn't make as much sense
19:55:05 <jbk> i mean, i'd like to improve it since it would make it easier to test certain things
19:56:41 <richlowe> Oh right, I'm just saying that logic is self-reinforcing
20:01:03 <jbk> just i've got several other things in front of it
20:23:35 <josephholsten> what's the deal with DEFAULTINURI in https://github.com/illumos/illumos-gate/blob/master/usr/src/tools/scripts/onu.sh.in#L32?  I feel like I'm doing something wrong just following https://illumos.org/docs/developers/build/#installing-your-build and getting an error about ips.sfbay repo
20:33:26 <richlowe> you forgot to specify -d probably?
20:33:49 <richlowe> ipkg.sfbay is the ancient ancient default pkg server inside sun microsystems, that nobody fixed because there's no... default anymore?
20:33:57 <richlowe> we should probably make it fail usefully, rather than trying to look that up
20:57:44 <jclulow> concur!
21:40:44 <sommerfeld> I'll fix it.  Filing a bug now..
21:41:05 <sommerfeld> https://www.illumos.org/issues/16799
21:41:05 <fenix> → BUG 16799: onu should not have a default package URI (New)
22:13:45 <jclulow> sommerfeld: thank you!