[06:12] <alkisg> vorlon: great, except that... then all shims need to add the distro name in the paths that they use, otherwise they end up in the same grub name. They don't currently do that, afaik.
[06:13] <alkisg> So it would be: UEFI > ubuntu shim > ubuntu grub > fedora shim > meh, ubuntu grub again, as the shimx64.efi name is hardcoded for all distros, with no distro name in their paths
[06:14] <alkisg> (same for dhcp/tftp too)
[06:25] <vorlon> alkisg: oboting ubuntu grub -> fedora shim means booting the fedora shim by its full path; and shim loads grub from the same directory
[06:25] <vorlon> booting
[06:26] <alkisg> vorlon: ah, thanks; I'll run some tests, as I think that many paths there were hardcoded, but I'm not sure if it was in the pxe or the local disk case. Merci!
[06:27] <voltagex> Hi, does anyone know what the performance impacts of ZFS on kernel  5.0 will be? Looks like a pretty major patch. I'm on an Atom 2758 box with 32TB of disks and I don't see another way of testing this short of running a full scrub (20 hours or so)
[07:03] <Skuggen> ahasenack: my_load_defaults should be there, but not load_defaults
[07:03] <Skuggen> The net-snmp package I tested with some time back was patched to add the my_
[10:17] <ahasenack> Skuggen: yeah, but it didn't find my_load_defaults. I wonder if it's related to the missing my_global.h header file
[10:18] <ahasenack> Skuggen: net-snmp upstream has basically an if/then/else now. It tries load_defaults, then my_load_defaults, and if neither exist, uses mysql_options()
[10:25] <ahasenack> Skuggen: line 450: https://sourceforge.net/p/net-snmp/code/ci/master/tree/apps/snmptrapd_sql.c
[10:25] <ahasenack> it calls mysql_options (if conditions are met) in line 565
[10:25] <ahasenack> I'm trying to backport that
[10:43] <Skuggen> ahasenack: my_global (and my_sys) have been removed from 8.0. Should only need mysql.h
[10:44] <ahasenack> Skuggen: and my_load_defaults? A grep for it in /usr/include -r doesn't find it
[10:47] <Skuggen> Hm, maybe I was wrong
[10:47] <ahasenack> it looks like mysql_options() can be used to load ~/.my.cnf
[10:48] <ahasenack> I'll keep on that path
[10:50] <Skuggen> Yeah, looks like that's the intended way now https://dev.mysql.com/doc/refman/8.0/en/mysql-options.html
[10:52] <ahasenack> k
[11:01] <Skuggen> Talked with devs. my_load_defaults was added back to the export list when we did the 5.7 transition in 16.04, but it was only meant to be a temporary measure :)
[11:04] <ahasenack> I see that kind of thing happening a lot
[11:05] <ahasenack> found a bug in mysql8 where another name, that had been removed, was added back by mistake (in the 8 series)
[11:05] <ahasenack> forgot what it was
[11:05] <ahasenack> so the bug was to remove it again, for good :)
[11:32] <rbasak> Has anyone successfully used "git ubuntu import-ppa" recently?
[11:50] <ahasenack> for cacti-spine: https://bugs.launchpad.net/ubuntu/+source/cacti-spine/+bug/1814526
[11:50] <ahasenack> rbasak: I didn't even know it existed
[11:50] <ahasenack> rbasak: do you have a tag for mysql8 ftbfs?
[11:56] <rbasak> ahasenack: I invented mysql-8.0-transition earlier. Perhaps use that and the ftbfs tag?
[11:57] <rbasak> Skuggen: FYI ^
[11:57] <Skuggen> ahasenack: I'm testing just putting my_bool back in now (it's a meaningless type, but at this stage it's better to patch mysql than the ~50 packages that use the type :))
[11:59] <Skuggen> Can use cacti-spine to verify the fix
[12:00] <ahasenack> ok
[12:01] <ahasenack> we could still use the bugs, though, to slowly migrate the packages away from my_bool then
[12:02] <Skuggen> Yeah. I sent in bug reports about this for all rdeps of libmysqlclient21 (or I thought I did, but maybe the commands I used to find them didn't list everything)
[12:03] <ahasenack> bugs in ubuntu?
[12:03] <ahasenack> any particular tag or pattern to search for?
[12:04] <Skuggen> "FTBFS with MySQL 8.0"
[12:04] <Skuggen> A few of them are for other things like my_sys.h and/or my_global.h, but the my_bool issue is the most common
[12:06] <ahasenack> I hit 4 in net-snmp
[12:06] <ahasenack> - my_bool
[12:06] <ahasenack> - my_load_defaults
[12:06] <ahasenack> - my_glpobal.h and my_sys.h
[12:06] <ahasenack> - something with my_init, MY_INIT, but that might have changed in a later 5.7 already
[12:06] <Skuggen> I got net-snmp to build by fixing the two headers and my_bool, I think
[12:06] <ahasenack> my_load_defaults is gone, that needs fixing
[12:07] <ahasenack> it's the last fix I need
[12:07] <Skuggen> But net-snmp can use mysql_options, can't it?
[12:07] <ahasenack> yes
[12:07] <ahasenack> that is the fix
[12:07] <ahasenack> it's just not in our package yet
[12:08] <ahasenack> about init, it was "there is mysql_init(), MY_INIT(), and my_init()"
[12:08] <ahasenack> if-defs all over the place
[12:09] <Skuggen> Maybe my_load_defaults was still in there in earlier 8.0 versions (it was an "unofficial" export in 5.7) and only removed recently.
[12:09] <Skuggen> https://launchpad.net/~mysql-ubuntu/+archive/ubuntu/mysql-8.0/+packages test of cacti-spine
[14:08]  * apw just did a dist-upgrade in disco ... and grub-efi-amd64-signed and shim-signed failed to install
[14:11] <apw> which seems to be stemming from grub-efi-amd64-signed complaining about 'db is empty' maybe
[14:24] <apw> /usr/share/grub/grub-check-signatures <-- seems it is this which is failing
[15:24] <apw> it seems this is because mokutil --export --db is saying db is empty; and this is fatal to bootloader installation
[15:53] <ahasenack> hi, does anybody know what this means: ERROR: erroneous package: rules extract failed with exit code 1
[15:53] <ahasenack> https://pastebin.ubuntu.com/p/JxJFfCVB24/ more context
[15:53] <ahasenack> it's an autopkgtest run in disco, for freeipa
[15:53] <ahasenack> freeipa is in disco-proposed
[15:53] <ahasenack> but apparently was removed from disco, is that it?
[15:54] <ahasenack> hm, looks like freeipa needs to be removed from disco-proposed then?
[15:54] <ahasenack> https://bugs.launchpad.net/ubuntu/+source/dogtag-pki/+bug/1813155
[15:54] <ahasenack> tjaalton: ^ ?
[17:00] <tjaalton> ahasenack: I'll upload a client-only version soon
[17:01] <tjaalton> to debian first
[18:20] <LocutusOfBorg> tjaalton, hello, what is the rationale for this delta? https://launchpad.net/ubuntu/+source/httpcomponents-client/4.5.6-1ubuntu1 ? can it be syncd?
[18:21] <tjaalton> LocutusOfBorg: yep
[18:33] <LocutusOfBorg> tjaalton, can you please sync it, or should I do it?
[19:15] <tjaalton> LocutusOfBorg: done
[23:13] <jilocasin> evening all
[23:13] <jilocasin> anyone have any idea why *the* default installer for 18.04.1 LTS Server is a subiquity based cloud one?
[23:16] <rbasak> I answered you in #ubuntu-server.
[23:17] <rbasak> !crosspost
[23:17] <jilocasin> rbasak: thanks, I responded there as well.
[23:18] <jilocasin> rbasak: sorry, I was told in #ubuntu to ask in #ubuntu-devel, #ubuntu-discuss, snd #ubuntu-server
[23:21] <rbasak> Sorry you were misled. General consensus on Ubuntu IRC channels is what the bot says on that - please don't crosspost.
[23:26] <krytarik> They didn't indicate to post it in all of them at once though.