02:53:26 <tozhu> hello all, a simple question: for Intel CPU, there are base frequency and turbo frequency, and the frequency is very high; my question is how to speed up the CPU on Turbo frequency in illumos/smartos? is it possible? and how to do if it’s okay, is there anyone tried?
07:56:06 <tomww> shouldn't that happen automatically by BIOS Settings? You may watch with powertop
10:02:39 <tozhu> tomww: Thank you
10:02:47 <tozhu> I’ll take a loo
10:02:51 <tozhu> look
15:24:28 <Aedil> I wish you all good progress in your doings. Excel!
15:46:31 <gitomat> [illumos-gate] 16331 loader: return errors from writing ZFS labels -- Warner Losh <imp⊙Fo>
16:13:09 <sommerfeld> found the paper on the collision in reduced-round SHA-256 discussed here yesterday: https://eprint.iacr.org/2024/349.pdf
16:15:56 <sommerfeld> what's new is they had a more efficient search for the 31-round collision (reduced from 2^65.5 to 2^49.5).   full sha-256 is 64 rounds.
18:49:38 <jclulow> papertigers: Sorry, I've sorted out the indexing job again
21:08:11 <jbk> just to ask (because I'll probably end up doing this in the next few weeks if not), anyone happen to be planning to add mhd(4I) support to nvme any time soon?
21:43:54 <nomad> Quick question: if you have two different sized drives in a zpool (e.g 12TB & 20TB) and you have one hot spare of each size is there any way to make sure that a failing 12TB drive is replaced with the 12TB hot spare?
21:44:02 <nomad> We've had a few situations now where the system grabs the larger hot spare to replace a smaller drive.
21:44:38 <nomad> I asked about this on #zfs and they suggested it would require a code change in cmd/zed/agents which comes from illumos according to the README.md in there.
21:44:59 <nomad> The situations I'm looking at right now are all in OmniOS, not ZoL.
22:00:20 <jbk> zed is very much an openzfs thing and not illumos
22:00:54 <nomad> understood. I'm just going by the README.md in there which says the code is lifted from illumos.
22:01:30 <nomad> but that's a side quest. My main thing is trying to get our omnios hosts to grab the correct hotspare.
22:02:37 * nomad does not use ZoL in production anywhere, but he does have multiple OmniOS ZFS fileservers.
22:38:50 <sommerfeld> nomad: you may be looking for the syseventd plugin in usr/src/cmd/syseventd/modules/zfs_mod 
22:39:32 <sommerfeld> (or not?)
22:55:32 <sommerfeld> strike that.   you want to look at usr/src/cmd/fm/modules/common/zfs-retire/zfs_retire.c
22:55:37 <sommerfeld> the function replace_with_spare
22:55:57 <sommerfeld> "Try to replace each spare, ending when we successfully replace it."
22:57:35 <sommerfeld> likely want to sort the list of spares by size and start with the smallest one that's big enough.
23:02:41 <nomad> that's great but I'm not a coder.
23:03:01 <nomad> I suppose I could put in a ticket. It seems kind of an obvious thing to me that you would want to replace like with like.
23:24:57 <sommerfeld> nomad: so, it's possible that you could get the behavior you want by ensuring that the spares are listed in order from smallest to largest.
23:36:17 <sommerfeld> nomad: and I agree -- it's a reasonable request.   But then there's the question of how far you go in terms of picking the spare that is most "like" the disk that failed.
23:37:37 <sommerfeld> do you just look at size, or at other parameters (SSD vs spinny disk.  If spinny, disk RPM?  manufacturer?  )
23:37:59 <sommerfeld> or disk model?