14:33:56 has anyone messed with the fault injection bits in sd ? 14:34:56 (unfortunately, I don't think any tools to utilize it are in illumos-gate) 14:51:13 i used it and previously google would return a search result linking to a developer's blog along with a small utility code using those ioctls 14:51:23 don't seem to be able to find it at the moment 14:52:50 i want to be able to fail a certain # of I/Os to a disk w/ a given key/asc/ascq.. from what i've seen so far it seems like like it's geared to inject actual scsi pkts into the mix 14:52:56 but wasn't 100% sure 15:00:42 (i can obviously add it, but if there's a way to do it with what's there that's just not obvious from first glance, i'd rather utilize the existing bits) 16:49:37 The illumos Developer's Guide states "While Sun made almost all of the operating system open source, there are still a few closed components that the illumos community is replacing." To what degree is this still true today? 16:55:16 there's still a few things, though mostly stuff people don't use anymore 16:56:39 jbk Thanks for your response. So is it safe to say "nothing essential"? 17:01:42 maaaybe the cpu_ms modules 17:02:17 though I'm not sure if they're 100% essential vs. just provide some extra bits (never looked at them too closely tbh) 17:07:06 :-) I'm new sorry, what do the cpu_ms [kernel] modules do? 17:08:03 (Roughly speaking) 17:09:12 rmustacc would know better, but I think mostly handle machine check exceptions and generate fmd ereports from them... there might be some other bits for specific CPU models as well, but that's my guess at least 17:12:23 I'm curious but I think that's enough detail for today. 17:13:16 :-) I'm sure these components can/will be replaced if the need arises? 17:13:57 as much as possible 17:14:21 i suspect some of the drivers for things like ancient NICs or such will probably just go away eventually 17:14:33 just because no one is using them 17:14:42 so there's no real incentive to replace them 17:17:56 That makes sense, thanks. I guess what I'm really curious about/trying to establish is how "healthy" illumos is, as I start to put more time into it. 17:18:39 i mean, i don't think i've run into an issue where any of those have posed a problem or somehow hindered anything 17:18:48 mostly it's about getting drivers written for new hardware :) 17:20:07 :-) that old chestnut... 17:28:15 * nomad mutters about drivers for new hardware 17:28:38 I'd *love* to see some progress on https://www.illumos.org/issues/15302 for example. 17:28:39 → BUG 15302: broadcom 3108 kernel driver update (New) 17:28:42 * nomad goes back to muttering in the corner 17:32:39 which reminds me, I need to see if support for the Broadcom HBA 9500 ever made it into OmniOS. 17:32:45 * nomad adds yet another ticket to the list. 20:36:07 what's the pci id? 20:49:16 jbk: I don't have either card installed in a host right now so can't query it. 20:49:34 I'd like to think I put it in the ticket (if you're talking about the 3108) but I probably failed to do that. 21:42:02 i mean the 9500 21:42:22 it's the least ambiguous and confusing way to identify the card 21:42:44 and the easiest way to tell with a reasonable amount of certainty 21:42:55 drivers go by the pci/pciex id the card presents 21:43:36 there is the lmrc driver that's out for review that adds some, but which ones those translate to in terms of marketing names/project names/ or whatever other names they get called is another manner 21:43:48 * nomad nods 21:44:01 Next time I'm in the office where I have it I'll try to find a computer I can plug it into to check.