11:55:03 <nahamu> richlowe: I don't check IRC as often as I used to. Tailscale seems generally status quo. I do rebases as they release new stuff, but I haven't done any real work in it otherwise.
11:56:19 <nahamu> Also I often forget to post an announcements and since I just use the binaries I've been bad about maintaining the packages, but my fork gets pretty regular tags and corresponding binary releases.
11:57:29 <nahamu> https://github.com/nshalman/tailscale/releases is the place to check.
18:41:49 <papertigers> fwiw I have been pulling those releases and everything seems to function properly 
19:52:33 <tsoome> hm... sata disks... 642M repaired, 56,97% done, 0 days 02:22:27 to go
19:55:23 <jbk> sata or ata?
19:56:13 <tsoome> sata
19:56:54 <jbk> our ATA driver is painfully slow (which is why anyone using gen 1 hyperv VMs might have thought they were running off tape :P)
19:57:03 <tsoome> WDC-WD4004FZWX and HGST-HUS726T4TALA6L4
19:57:21 <jbk> wonder if it's maybe the model
19:58:04 <jbk> (drive model)
19:58:14 <jbk> my HGST drives i don't think are that slow resilvering
19:58:45 <jbk> and *knock on wood* have been pretty error free
19:59:16 <jbk> which is more than i can say of a certain model of SAS drives
19:59:24 <tsoome> this supermicro X10SAE is pretty old
20:05:08 <jbk> though speaking of that, i'll hopefully have some sata improvements up for review soon..
20:05:31 <jbk> trying to test those, hit a different bug that made it harder to test :)
20:08:39 <tsoome> :D
20:11:31 <sommerfeld> resilver/scrub speed depends a lot on the "shape" of data in the pool.  Lots of tiny files vs. big sequentially written files.
20:26:48 <tsoome> thats true, of course.
21:08:23 <richlowe> an omnios used filed #16590 (fenix) which seems intriguing, if you're networky
21:08:24 <fenix> BUG 16590: speed difference between IPv4 & IPv6 on 10Gbit connections  (New)
21:08:25 <fenix> ↳ https://www.illumos.org/issues/16590
21:10:06 <tsoome> uh. ouch.
21:12:47 <nomad> I'm happy to provide additional data/tests as directed.
21:23:37 <jclulow> sommerfeld: I feel like at least with the sequential resilvering stuff that's less of an issue these days