06:02:37 Hi illumos. 08:30:01 Hi! 08:57:43 Good day. 17:18:46 hey there, I'm trying to setup a point-to-point link between my OmniOS server and my router using a /31 subnet, 17:19:17 and I get "ipadm: Could not create address: Can't assign requested address" 17:20:10 say I'm trying to use the 100.100.100.100/31 pair, my command is as follows 17:20:51 pfexec ipadm create-addr -T static -a 100.100.100.101/31 bge1/v4 17:21:15 could it be that OmniOS does not support /31 subnets? I was using a /26 with no issue before that 17:28:26 also tried the following: 17:28:31 pfexec ipadm create-addr -T static -a local=100.100.100.101,remote=100.100.100.100 bge1/v4 17:28:56 which outputs "ipadm: Could not create address: Invalid argument provided" 17:36:09 I'm not sure many ethernet systems will like a /31 mask, you might want a /30 giving 4 addresses (100 is netnum, 101, 102 are useable, 103 is broadcast) 17:38:43 well, RFC 3021 is already 25 years old, 17:39:21 It's an ipadm bug: https://www.illumos.org/issues/13232 17:39:22 → BUG 13232: Add support for IPv4/31 or IPv6/127 (New) 17:39:29 The kernel and ohters support using /31s. 17:39:29 and I'm using point-to-point links on FreeBSD & Linux with no issue, so I assumed Illumos would support it as well 17:39:41 oh, thanks rmustacc 17:40:55 Yeah, point-to-point IPv4 stuff works fine once you get it set up. 17:41:00 it looks like you just need to use a different syntax 17:41:06 to workaround the bug 17:41:14 Unfortunately, yes. 17:44:36 thanks yall, the syntax is a little convoluted and of course I fucked it up on first try, lol 18:09:46 will the "ifconfig 100.100.100.101 destination 100.100.100.100 netmask 255.255.255.255 up" resist rebooting the machine? 18:13:14 manually issued (command line) ifconfig is not persistant, if that's what you are asking. 18:25:07 indeed, that was my question, since I cannot use ipadm for that specific scenario 18:25:19 ipadm does show the point-to-point link now though 18:25:43 where shall I add the custom ifconfig command so it survives a reboot, may I ask? 18:26:09 I'm obviously new to Illumos, migrating away from Linux for VM management 18:28:55 from earlier did you try 'ipadm create-addr -T static -a local=100.100.100.101 -a remote=100.100.100.100 bge1/v4' (two separate -a options, not -a xxx,yyy) ? 18:28:57 veg: bug 13232 linked above suggests trying: ipadm create-addr -T static -a local=192.168.1.5 -a remote=192.168.1.4 e1000g0/v4p2p 18:29:18 which would be persistent 18:29:30 yes, I did try the 2 -a options, with no luck 18:29:42 (also, if you already have bge4/v4 address, that will generate an error) 18:30:23 the bug report also had the local & remote addresses reversed in the ifconfig example 18:31:22 -a local -a remote & -a local=xx,remote=xx give different error messages but neither works, unfortunately 18:32:06 does ipadm show an existing bge1/v4 entry? 18:32:48 no, there was none 18:39:45 If you are testing after issuing the ifconfig then it will fail because the device is configured. You have to clear the config before trying new ipadm commands. 19:41:37 I ha/clean 19:41:43 woops 19:43:28 oi 19:45:36 looks like other people had similar experiences with the ipadm command, nomad : https://www.reddit.com/r/illumos/comments/y53yy0/rfc_3021_31_network_support/ 19:46:47 today I learned there's an illumos reddit? 19:49:30 it's been there for a while... 19:49:31 Yeah, I keep an occasional eye on it and try to dampen any misinformation 19:58:17 there's even a discord too (if papertigers ever gets around to finishing it :P) 20:00:36 gosh 20:17:47 there's a reddit for everything! :D 21:41:59 [illumos-gate] 11989 DHCP client support for "Interface MTU" option -- Joshua M. Clulow 22:19:21 is there a dcmd that'll show you how long a sleeping kernel thread has been sleeping? 22:19:56 or would maybe comparing kthread_t->t_hrtime to 'now' on a sleeping kthread show that? 22:22:20 jbk: I think "::thread -d" gives you the dispatcher state, including the SWITCH column which seems like a time 22:22:49 might be worth looking at how that works 22:24:31 the comment in thread.h seems like enough? 22:25:28 the SWITCH column is ... different 22:27:22 I assume that's the low-res version of what the hrtime is hr of 22:28:20 ah 23:52:39 hah 23:52:52 figures there's an XXX command for the exact thing I was looking at 23:54:18 https://github.com/illumos/illumos-gate/blob/master/usr/src/cmd/mdb/common/modules/zfs/zfs.c#L4066-L4072