July 19, 2012 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
July 19, 2012 [00:31:58] *** haggle has quit IRC[00:35:25] *** knweiss has quit IRC[00:38:58] *** dap has joined #smartos[01:01:26] *** Zigara has joined #smartos[01:03:29] *** Licenser has quit IRC[01:04:01] *** dap has quit IRC[01:04:32] *** Licenser has joined #smartos[01:06:40] <Zigara> I have a supermicro server with a SCU port, I attached a mini sas cable and plugged in more drives, they show up fine under linux, but under smartos it will not even show the scu ports[01:06:44] <Zigara> any ideas on how to go about this?[01:08:40] <ryancnelson> is that attached to a raid controller?[01:09:14] <Zigara> yes, I'm able to disable the raid part in the BIOS though[01:10:33] <ryancnelson> but it's still on that controller, which i'd not be surprised to find out we don't have a driver for[01:10:47] <ryancnelson> … can you be more specific about your hardware?[01:10:59] <Zigara> http://www.supermicro.com/products/system/2u/6027/sys-6027r-trf.cfm?parts=show[01:12:36] <rmustacc> SCU is not supported.[01:12:40] <rmustacc> Sorry.[01:12:45] <Zigara> bleh[01:12:45] <rmustacc> It's not on the HCL.[01:13:01] <Zigara> this barebone came with 8 hotswap bays but only 6 normal sata ports[01:19:53] <wesolows> if you are doing business with SuperMicro, get them to offer you the SAS2308 in IT mode in that system.[01:20:23] <wesolows> They offer such options in most of their boards and chassis, including ones that are extremely similar to the 6027R series.[01:20:42] <wesolows> The SAS2308 in IT mode is 1000,86; it is not in the HCL but it works.[01:20:59] <wesolows> (where "works" means it sees drives and does I/O; do your own homework if you care more)[01:21:28] <Zigara> thanks for the advice, I appreciate it[01:32:12] *** jelmd has quit IRC[01:38:35] *** jelmd has joined #smartos[02:21:04] *** ryancnelson has quit IRC[02:24:39] *** dap has joined #smartos[02:28:41] *** darjeeling has quit IRC[02:36:33] *** relling has quit IRC[02:36:47] *** relling has joined #smartos[02:59:07] *** denizr1 has joined #smartos[03:00:09] *** denizr has quit IRC[03:15:07] *** darjeeling has joined #smartos[03:17:23] *** jdavis has quit IRC[03:50:45] *** noahmehl has joined #smartos[03:50:54] *** darjeeling has quit IRC[03:51:41] *** denizr1 has quit IRC[03:56:51] *** enmand has quit IRC[04:10:51] *** denizr has joined #smartos[04:17:41] *** darjeeling has joined #smartos[05:18:09] *** aroop has joined #smartos[05:18:22] *** denizr has quit IRC[05:36:31] *** noahmehl_ has joined #smartos[05:36:32] *** noahmehl has quit IRC[05:36:33] *** noahmehl_ is now known as noahmehl[05:37:40] *** Webhostbudd has joined #smartos[05:39:18] *** noahmehl has quit IRC[05:39:19] *** noahmehl_ has joined #smartos[05:39:19] *** noahmehl_ is now known as noahmehl[05:40:25] *** noahmehl_ has joined #smartos[05:40:25] *** noahmehl has quit IRC[05:40:26] *** noahmehl_ is now known as noahmehl[05:40:58] *** noahmehl_ has joined #smartos[05:40:58] *** noahmehl has quit IRC[05:40:58] *** noahmehl_ is now known as noahmehl[05:48:49] *** ryancnelson has joined #smartos[06:14:33] *** dap has quit IRC[07:20:36] *** marsell has quit IRC[07:22:04] *** aroop has quit IRC[07:26:10] *** aroop has joined #smartos[07:32:23] *** jelmd has quit IRC[07:51:52] *** deirdres has quit IRC[07:58:30] *** deirdres has joined #smartos[08:00:18] *** ryancnelson has quit IRC[08:08:07] *** max` has joined #smartos[08:08:11] *** deirdres has quit IRC[08:13:49] *** mathie has joined #smartos[08:25:19] *** Licenser has quit IRC[08:25:56] *** jamesd has quit IRC[08:48:18] *** bens1 has joined #smartos[08:48:28] *** mathie has quit IRC[09:12:13] *** Licenser has joined #smartos[09:12:56] *** noahmehl has quit IRC[10:02:32] *** Licenser has quit IRC[10:04:10] *** Licenser has joined #smartos[10:12:16] *** marsell has joined #smartos[10:21:12] *** darjeeling has quit IRC[10:21:19] *** darjeeling has joined #smartos[11:02:52] *** pjama has joined #smartos[11:37:47] *** darjeeling has quit IRC[12:43:00] *** sbfox has joined #smartos[12:43:22] <sbfox> Anybody alive in here?[12:43:49] <Licenser> nope we're just zombies[12:44:30] <sbfox> Cool :)[12:45:05] <sbfox> I have an imgadm question, I cant find out how make imgadm work thru a proxy[12:45:29] <sbfox> ive tried setting environment variables[12:46:20] <Licenser> did you tried uppercase and lowercase env wars?[12:46:23] <Licenser> they are a bit icky at times[12:46:57] <sbfox> only lowercase so far, hold on till try upper[12:47:31] <Licenser> just an idea I don't like proxies they are so 1990[12:48:00] <sbfox> I dont have a choice here[12:48:47] <sbfox> looks like a no[12:52:29] *** mamash has joined #smartos[12:53:32] <Licenser> yea I know that[12:53:40] <Licenser> what you can do is use CURL to get the data and import it manually[12:54:00] <Licenser> or use something like nomnom to set up a local verion of the repository[12:55:12] <sbfox> I think curl and local mirror[12:55:32] <Licenser> *nods*[12:55:42] <Licenser> look at nomnom if you want or set up your own stuff it's not that hard[12:56:06] <sbfox> Cheers[12:56:29] <MerlinDMC> Licenser, imgadm uses curl to get the dataset stuff ;)[12:56:48] <Licenser> I don't think so it uses a library perhaps but not curl[12:56:55] <Licenser> curl should repsect the proxy env vars[12:57:02] <MerlinDMC> but i think the first request for the json is not done with curl[12:57:38] <sbfox> Where is the datasets json file stored locally?[12:57:53] <MerlinDMC> sbfox, /var/db/imgadm/imgcache.json[12:58:06] <sbfox> Ta[13:02:01] <sbfox> Well the dataset retrieval worked, import throws an error "throw new TypeError("Parameter 'url' must be a string, not " + typeof url)"[13:02:19] <sbfox> I guess I could curl the image as well[13:03:55] <MerlinDMC> the image is retrived by using curl that should work[13:06:05] <sbfox> imgadm import e4cd7b9e-4330-11e1-81cf-3bb50a972bda[13:06:05] <sbfox> url.js:96 throw new TypeError("Parameter 'url' must be a string, not " + typeof url)[13:08:36] <MerlinDMC> what environment variables did you set?[13:09:23] <sbfox> http_proxy https_proxy[13:09:30] <sbfox> upper case[13:09:56] <sbfox> same with lower case[13:10:28] <sbfox> I can curl the image no problems[13:17:47] <MerlinDMC> Licenser, during `imgadm update` a _url property is added for every manifest ... after a simple curl that one is missing and forcing an exception[13:18:05] <MerlinDMC> there: https://github.com/joyent/smartos-live/blob/master/src/img/node_modules/imgadm.js#L645[13:23:49] *** darjeeling has joined #smartos[13:46:53] *** jelmd has joined #smartos[13:50:30] *** enmand has joined #smartos[14:36:15] *** LeftWing__ has joined #smartos[14:37:20] *** LeftWing has quit IRC[14:42:58] *** sbfox has quit IRC[14:55:34] *** aroop_ has joined #smartos[14:55:34] *** aroop has quit IRC[15:02:28] *** LeftWing__ is now known as LeftWing[15:39:06] *** darjeeling has quit IRC[16:08:12] *** joshie_ is now known as joshie[16:08:47] *** noahmehl has joined #smartos[16:12:21] *** noahmehl has quit IRC[16:20:29] *** enmand has quit IRC[16:46:12] *** mamash has left #smartos[16:56:05] *** jamesd has joined #smartos[16:58:03] *** jdavis has joined #smartos[17:00:32] *** darjeeling has joined #smartos[17:04:48] *** denizr has joined #smartos[17:07:49] *** deirdres has joined #smartos[17:08:43] *** kfr has joined #smartos[17:10:15] <kfr> getting a cron job email "sh: line 1: /smartdc/bin/sdc-lastcomm: not found" on joyent_20120713T005832Z[17:10:23] <kfr> should I create a bug?[17:10:51] *** d[^_^]b has quit IRC[17:11:00] *** d[^_^]b has joined #smartos[17:17:50] <jeffpc> I have a KVM guest that I want to move to smartos, it should just be a matter of replacing the dataset after vmadm create, right?[17:30:27] <rmustacc> kfr: Sure.[17:30:41] <rmustacc> jeffpc: Should be.[17:41:02] <jeffpc> ok, I think tonight is the night my server will go smart :)[17:45:14] *** bens1 has quit IRC[17:46:30] *** mamash has joined #smartos[17:47:04] <aroop_> in smartos, how do i see , which all ports are listening...[17:47:13] *** bens1 has joined #smartos[17:47:23] *** bens1 has quit IRC[17:47:28] <nahamu> aroop_: "netstat -an" perhaps?[17:47:54] <nahamu> I haven't figured out how to filter on just tcp[17:48:11] *** jerry__ has joined #smartos[17:48:32] <rmustacc> jerry__: ping[17:48:36] <aroop_> yes...got it.. ..ty[17:48:38] *** jerry__ has quit IRC[17:52:56] *** jerry__ has joined #smartos[17:56:47] *** enmand has joined #smartos[17:59:38] *** jerry__ has left #smartos[18:01:42] *** aroop_ has quit IRC[18:03:47] *** dap has joined #smartos[18:05:27] *** ryancnelson has joined #smartos[18:11:00] *** jdavis has quit IRC[18:19:18] *** aroop has joined #smartos[18:32:19] *** bens1 has joined #smartos[18:43:08] *** aroop has quit IRC[18:45:16] *** ryancnelson has quit IRC[18:45:40] *** ryancnelson has joined #smartos[18:46:14] *** aroop has joined #smartos[18:46:37] *** dap has quit IRC[18:52:12] *** robg has joined #smartos[18:53:43] <nahamu> yay, it's robg![18:56:03] <robg> I figured it was about time I joined office hours ;)[18:59:18] <joshie> joy[19:00:17] *** max` has quit IRC[19:00:40] <wesolows> good news, everyone! it's SmartOS office hours time again[19:02:20] <bdha> Yay![19:02:46] <richlowe> g'afternoon bdha[19:04:24] <joshie> what is the status of ipv6 stuff? :D[19:05:16] <robg> joshie: I'm just looking at your changes now. Looking good so far.[19:05:48] <joshie> I am not a developer by trade so feel free to laugh[19:09:21] <nahamu> who would be in charge of joshie's listen_ip and allow_conflicting_ip features?[19:10:01] <joshie> I changed the feature name to allow_ip_conflict but not the branch name[19:10:14] <joshie> sounded mouthy :)[19:11:14] *** aroop has quit IRC[19:11:16] *** aroop_ has joined #smartos[19:12:24] *** ryancnelson has quit IRC[19:12:52] *** deirdres has quit IRC[19:13:03] <rmustacc> nahamu: Josh is the final reviewer on all vmadm changes.[19:13:19] <rmustacc> joshie: At the least you're going to need to make sure that you have manual page updates and tests where appropriate[19:14:24] <joshie> I figured as such. I am more curious if such features would even be accepted before doing that.[19:15:07] *** trentm has joined #smartos[19:15:25] <joshie> What kind of tests?[19:15:35] <killfill> Maybe its too boring, and has nothing to do with office-hourse.. :) but im still trying to understand cpu-shares... why would i get a different result than Jerry (from the mailinglist)?[19:20:26] *** jdavis has joined #smartos[19:21:07] <nahamu> is there documentation for how to run the existing tests?[19:22:12] *** mamash has left #smartos[19:22:32] <rmustacc> nahamu: Probably not.[19:23:55] *** trentm has quit IRC[19:24:44] *** aroop_ has quit IRC[19:24:49] *** aroop has joined #smartos[19:24:52] <robg> The docs for running vmadm tests are here: https://github.com/joyent/smartos-live/blob/master/src/vm/README.md[19:25:23] <robg> (I'm pretty sure that's up to date)[19:28:55] *** Licenser has quit IRC[19:31:01] <joshie> neato, I will do that after my changes from now on[19:33:38] <nahamu> do those tests need to be run on a GZ running the latest vmadm, or does it use its own local copy?[19:34:21] <siezer> function append_if_no_such_line() { if ! grep "`printf '%q' "$1"`" $2 ; then echo "$1" >> $2 ; fi }[19:34:23] <siezer> er[19:34:30] <siezer> (wrong chan sorry folks)[19:35:28] <robg> nahamu: The tests need to be run in the GZ, and they will run against the current vmadm.[19:36:10] <robg> (eg: https://github.com/joyent/smartos-live/blob/master/src/vm/tests/test-50-creates.js#L9 - it uses /usr/vm/node_modules/VM.js directly)[19:39:31] <joshie> robg: is the way I am using the hostname{,6} files an acceptable way for me to continue?[19:44:26] <joshie> tbh if I were designing it without limitations, I'd have a single hostname6 file per interface with a list of addresses and would use addif, to avoid conflicts (especially coupled with autoconfiguration), but that seems to fly in the face of the way solaris would treat those files[19:44:37] <nahamu> okay, so the workflow would be 1. write patches 2. rebuild smartos with patches applied and rebuild thet test suite 3. boot a machine on the fresh smartos and run the fresh test suite[19:45:13] <nahamu> do the tests create KVM branded zones?[19:49:39] <robg> joshie: having hostname6 files matches upstream illumos, so that seems fine to me. I'd rather we not surprise people too much by changing how those files work.[19:50:40] <joshie> then I will scratch the second idea despite it being easier :)[19:51:02] <joshie> or scrap it[19:53:12] <robg> nahamu: Yes - https://github.com/joyent/smartos-live/blob/master/src/vm/tests/test-create-kvm.js[19:54:22] <joshie> My final question would be, are these ip6 patches as is worthwhile to joyent enough for me to submit a pull request?[19:57:31] <robg> joshie: I think they're worthwhile, but I'd like to see a couple of things first before pulling them in.[19:58:21] <robg> 1) an update to the vmadm man page (particularly explaining that if you use cidr notation, you don't need netmask)[20:00:10] <robg> 2) I'm not sure about having the ip and ip6 properties be a comma-separated string in the vmadm payload. I'm thinking that we should make it an array (since it is JSON, after all).[20:00:40] <robg> And as rmustacc said above, Josh needs to take a look first.[20:01:32] <joshie> An array would make it easier for me to update the ip conflict code (which I had yet to do), so I am on board with that thought[20:02:22] <robg> Cool.[20:02:36] <joshie> Will it automatically treat a single string as an one-element array?[20:03:21] <joshie> If it were used in such a context, that is[20:04:40] <robg> We'll definitely want to maintain backward compatibility, so you can do "ip": "192.168.1.1" or "ip": [ "192.168.1.1" ][20:04:41] <rmustacc> JSON will not.[20:04:55] <rmustacc> So you'll have to manually do that.[20:04:56] <robg> But you'll have to detect if it's an array in the code and act accordingly.[20:05:24] <rmustacc> Which is as simple as Array.isArray("foo");[20:05:41] <rmustacc> Then you can transform it to the same underlying representation.[20:06:05] <rmustacc> If you are supporting multiple IPs, you need to cap that at the number that can be set in antispoof.[20:06:25] <nahamu> how many is that?[20:07:05] <rmustacc> I don't recall off hand.[20:07:11] <rmustacc> You'll have to dig into it and figure it out.[20:07:14] *** jamesd has quit IRC[20:07:24] *** jamesd_ has joined #smartos[20:07:54] <joshie> should I eliminate the ip6 field and detect the format of the ip on the ip field instead?[20:08:12] <joshie> I know as it stands vmadm doesn't validate if the ip field is a valid ip4 address[20:08:31] <rmustacc> That should be fixed then.[20:08:59] <joshie> because the total antispoof count would have to account for both ip4 and ip6 addresses[20:10:21] <joshie> (I had no idea that the anti spoof address list had an upper bound)[20:11:06] *** bens1 has quit IRC[20:12:54] <robg> joshie: the ip6 field seems reasonable to me. I'd rather they remain separate.[20:13:37] <joshie> Okay.[20:13:53] <joshie> I seem to be having trouble finding the maximum allowed-ips list amount[20:14:00] <joshie> but I could write a simple script to test this instead...[20:14:03] *** aroop has quit IRC[20:21:05] <joshie> 18 seems to be the number[20:21:07] <joshie> that's arbitrary[20:21:25] <rmustacc> You'd have to look in the kernel to figure out where it comes from.[20:23:01] <joshie> no wait[20:23:08] <joshie> it has to do with the length of the string of the option[20:24:27] <robg> joshie: Yeah, that sounds about right.[20:27:27] <joshie> ah, so it's the property list is limited to 255, and that includes the 'allowed-ips= part[20:29:57] <joshie> if only it allowed you to append...[20:35:01] *** deirdres has joined #smartos[20:45:17] *** robg has quit IRC[20:48:49] <nahamu> sounds like a flaw in dladm rather than the kernel struct, though I could imagine the kernel having a limit too.[20:49:43] *** Licenser has joined #smartos[20:57:59] <bahamas10> does anybody know how to update pkgsrc from joyent 2011q4 to 2012q1? i don't see an update.sh[20:58:20] <nahamu> mamash isn't around at the moment...[20:58:23] <nahamu> he would know.[20:58:51] <bahamas10> ok, does he usually idle in this channel[21:01:38] <nahamu> he's certainly here pretty regularly.[21:12:28] *** tafryn has quit IRC[21:14:25] *** mamash has joined #smartos[21:25:28] *** dap has joined #smartos[21:26:10] <nahamu> bahamas10: mamash is here.[21:26:20] <nahamu> mamash: < bahamas10> does anybody know how to update pkgsrc from joyent 2011q4 to 2012q1? i don't see an update.sh[21:28:32] <_jared_> is ulimit something that is limited by the global zone such that a project that sets a maximum doesn't actually work?[21:33:07] *** jamesd_ has quit IRC[21:36:20] <_jared_> is there a global zone quota mechanism?[21:36:35] <_jared_> besides the zfs disk quota[21:37:49] *** dap has quit IRC[21:38:30] <joshie> rmustacc: the string limit within the confines of set-linkprop is 256, and subtracting allowed-ip= its 243, which in terms of full ip6 addresses amounts to 6 ip's. You think it's better to limit the amount of combined ip4/6 addresses to 6 (something that makes sense to the common man), or limit the combined length of the ips to 243[21:38:38] *** dap has joined #smartos[21:38:48] *** dap has quit IRC[21:40:19] <joshie> (which in the ip4 world allows possibly up to 19 ips)[21:41:04] *** Z`` has joined #smartos[21:42:00] <rmustacc> I would opt for minimal surprise by default. Which I think means a common limit of 6 ips.[21:44:19] *** jmeredith has joined #smartos[21:44:48] <_jared_> rmustacc: thoughts on my question above, is ulimit something that needs to be set in the global zone?[21:45:19] <rmustacc> ulimit doesn't need to be set in the global zone per-se.[21:45:27] <_jared_> ok[21:45:45] <rmustacc> But the question of resource controls is a bit more complicated and I don't have the time to dig into the specifics currently, sorry.[21:46:07] <_jared_> np[21:46:09] <rmustacc> But if you're using projects, I would strongly suggest you read resource_controls(5) man page.[21:56:43] *** knweiss has joined #smartos[22:38:46] <ismell> I'm still running 2011Q3.. if i update to 2012Q1 will that break stuff ?[22:56:33] *** gregburd has quit IRC[22:58:58] *** Webhostbudd has quit IRC[23:06:28] *** bens1 has joined #smartos[23:16:41] *** knweiss has quit IRC[23:23:24] *** bens1 has quit IRC[23:29:41] *** denizr has quit IRC[23:33:00] *** denizr has joined #smartos[23:33:40] *** denizr has quit IRC