[00:28] Hi [00:28] Need help Nodes are not getting populated after successful booting through MAAS PXE Boot [00:44] can any one guide me to the URL or help please [07:40] Bug #1616772 opened: [2.1] Entering rescue mode reports successful events as failed [07:40] Bug #1616773 opened: [2.1] IP Address not automaticlly updated when in rescue mode [08:06] anyone who can help since i tagged vlan on my public interface on my nodes i cant deploy at all. keeps falling in the log it says that the environment is already boostrapped. I deleted manually .cloud-install and .juju and same thing tried all combinations of static/autoassign/unconfigured network settings and same issue tried to deploy 16.04 and 14.04 still stuck :( [08:10] upgraded maas to 2.0 maas server to 16.04 and added conjure-up that fails with cannot assign link-layer ip address found that it is logged as a bug gone back down to 14.04 on maas and 1.9.4 [08:11] is there a working combo that will deploy when using multiple vlans on the public interfaces on my nodes? [08:12] oz_: i dont understand exactly what you are trying to do [08:12] oz_: but maas can deploy machines with vlans [08:12] oz_: the first thing I'd recomend you do [08:13] oz_: is ensure you can deploy a machine iwthout juju [08:13] oz_: if that succeeds, then the issue you are experiencing is related to Juju [08:13] and I'd suggest you talk to someone in #juju [08:13] @roaksoax let me just give you quick overview [08:15] i have 10 nodes and they hve two interfaces one is maas interface separate network and the other is interface are plugged into vlan 10 and 101 which are added on the nodes since then i cannot deploy not with autopilot or with conjure-up [08:15] the switch ports carry multiple vlans tested [08:15] i will try with juju and see if there is any change [08:16] oz_: right, so I suggest you first try to deploy with maas [08:16] oz_: to try to identy whether the issue is maas or juju [08:17] i tried everything runs fine when i see the logs but when it gets to juju part it fails [08:18] @roaksoax i will try manually now and will et you know [11:09] Hi [11:10] How does upload commission script works? [11:13] And can we add our own action items in addition with power on, power off, Commission and so on? [12:09] sujeet_: the script gets stored in the db and then rendered when commissioning runs [12:09] sujeet_: you cannot change actions in MAAS [12:10] swhat actions were you looking to have ? [12:15] Hi [12:15] ls === plars_ is now known as plars === mup_ is now known as mup [13:53] sujeet_, ping? [15:10] kiko,roaksoax, did you see my messages about Centos7 issues? [15:10] prav, I did not, please repeat as my scrollback blew away [15:10] my browser crashed and I was not able to see anything [15:11] :-) [15:11] heh [15:12] cpl of issues actually. nodes are showing as "failed commissioning" and images trying to boot to centos7 says "boot-kernel-image" not found and stops at boot: [15:13] is there a good doc that shows how I can bring the nodes into prod.. enlist, commission,acquire, deploy etc? [15:18] prav: to enlist/commission you need ubuntu [15:19] one more thing, can I deploy a node with a "manual" power setting? [15:19] ah ok.. [15:21] prav: yes you can deploy with manual power setting (obviously, you need to power it on yourself and make sure it pxe boots) [15:22] so the initial part is handled by ubuntu and then the os deploy will be Centos7? [15:22] prav: ubuntu will deploy centos onto the disk [15:23] prav: a different way to put it is that the installer is ubuntu :( [15:23] :) [15:24] ok. [15:24] and it will install the od on the first volume it gets right? [15:24] "os" [15:25] prav: correct [15:25] ok.. I think I know how it works now, but why was I getting the image not found? The UI shows that it has imported Centos& [15:26] prav: exactly so when you deploy centos, MAAS boots into the Ubuntu ephemeral environment, downloads the CentOS image and installes it onto the disk [15:27] does it download from the node or from the maas server? [15:29] prav: from the maas server [15:30] hmm.. then I guess I need to check the repo file..or add a repo [15:30] prav: maas boots ephemeral environment, installer downloads the centos image from maas server, installer copies the image into disk, machine reboots onto centos [15:35] so the boot-sources read should tell me the location of the images [15:38] Bug #1616962 opened: [2.1] Additional repositories error surfacing uses old style [15:39] roaksoax: Hi roaksoax, i need the action item like firmware upgrade and secure erase using the storage card inserted on the PCI slot [15:41] roaksoax: Hi Andres Rodriguez, i need the action item like firmware upgrade and secure erase using the storage card inserted on the PCI slot [15:43] Hi Kiko [15:44] i need the action item like firmware upgrade and secure erase using the storage card inserted on the PCI slot [15:44] sujeet_, hmm, that's a somewhat involved process [15:44] firmware upgrading needs to be done via commissioning scripts [15:45] sujeet_, secure erase you just need to basically enable per-server [15:45] once that's done, it will wipe upon decommission [15:45] do you want to do it before deploying? [15:47] but i need to be doing our storage controller card [15:50] i am able to to see the storage controller card info on the xml/yaml after commissioning the node, but i need to do secure erase or firmware upgrade of drives and storage controller hardware [15:50] sujeet_, you mean the storage controller has a secure erase command you can run? [15:50] yes [15:50] sujeet_, not using MAAS' secure erase? [15:50] hmm [15:50] sujeet_, that is also a custom commissioning script [15:50] ok [15:51] sujeet_, you basically need to write commissioning scripts for your actions [15:51] those are run during commissioning [15:51] if you need help with those and want to have a commercial conversation just /msg me and I'll get you connected [15:52] how can we know that firmware upgrade is completed successfully on MAAS dashboard [15:52] ? [15:54] commercial conversation? i didn't get kiko, i am new to this webchat [15:55] sujeet_, if the firmware upgrade was successful, when you recommission the LSHW data will have been updated [15:55] sujeet_, I meant if you are interested in buying support or consulting to help you with your customization [15:57] ok Kiko [16:03] can we write sample script like displaying "hello" on MAAS? [16:05] sujeet_, well, yes, but rather than displaying hello, you can start with writing something and picking it up on the commissioning logs. [16:05] example? [16:08] sujeet_, can you have look at /etc/maas/commissioning* and see if there is anything? [16:09] sujeet_, or /etc/maas/templates/commissioning* [16:11] on "commissioning-user-dat"? [16:12] roaksoax,kiko: Thanks for your help so far! [16:32] sujeet_, yep [16:35] kiko : user_data_disk_erasing.template, user_data.template, user_data_config.template, snippets and user_data_poweroff.template [16:38] I'd not be relaying in those templates to do stuff [16:38] since thos will be soon removed [16:46] roaksoax, I think he just wanted a starting point [16:46] roaksoax, but scripts basically are run as roon inside the ephemeral image and can do anything, right? [16:46] kiko: pretty much [16:47] kiko: say, for example, if there's a hp utility for linux, you can have the script download that from the internet, and mess with your bios as a commissioning script [16:47] kiko: which is what we've actually done [16:47] yep [16:47] that's what I thought [16:51] and can we change the MAAS gui to show the info like hard drive for the storage controller? where i can see the info of storage controller in yaml/xml [16:54] sujeet_, are you using hardware raid? [16:54] sujeet_, we'd love to take patches that improve our support for hwraid controllers [16:55] if that's what you are looking at [16:55] yes [16:58] sujeet_, right, so the drives we show are the "logical" raided units [16:58] not the physical drives [16:58] we don't have a place to show the physical drives today [16:58] I'm not sure whether the XML output we generate has them, does it? [16:59] for now i have attached one drive to controller, and was able to see in xml [17:00] and also in the available disk section [17:01] the drive state is in raw [17:11] Hi kiko/roaksoax: can you remind me where I can make tweaks to the preseed stuff? I want to install "python-minimal" by default. I'm guessing /etc/maas/preseeds/curtin_userdata ? [17:13] shewless, http://askubuntu.com/questions/636837/are-there-examples-of-custom-installation-scripts/636867#636867 [17:22] kiko: thanks.. having a read now. [18:01] kiko: Do I have to restart any services for the changes to take effect? I tried deploying a server with my changes and they didn't seem to work [18:01] here is what I changed: http://paste.ubuntu.com/23089864/ [18:02] I added the python-minimal line [18:08] shewless, you should not need to restart anything [18:09] kiko: hmm. I don't see any errors in rsyslog.. not sure where else to look. any suggestions? Does it look like I added the line correctly and I should expect it to run on every new deployment? [18:10] roaksoax, I wonder if you can help shewless [18:15] hmmm.. maybe I added it to the if third_party_drivers and driver section.. which maybe isn't being hit [18:15] oooops [18:15] I will try and pull it out of that if statement.. not sure if that will solve it or not [18:17] shewless, I'm otp almost the whole day so kind of sucky at latency today [18:27] kiko: np. I fixed it by moving my install above the "if" statement in the late_commands block. We are good to go [18:28] shewless, woot [18:28] shewless, that's your /etc/maas/preseeds/curtin_userdata file? === mup_ is now known as mup [18:41] kiko: yes [18:42] cool === mup_ is now known as mup [19:47] Bug #1617044 opened: View Full History on Events doesn't show anything [20:02] Bug #1617044 changed: View Full History on Events doesn't show anything [20:05] Bug #1617044 opened: View Full History on Events doesn't show anything [20:15] shewless: i know why that didn't work [20:16] shewless: line 26 on your paste should be between line 17 and 18 [20:16] roaksoax: thanks I managed to figure that out. All working now! [20:18] cool [20:39] thanks roaksoax [20:49] Hi guys, did anyone tried to use mysql backend with maas? [20:49] osm: nope [20:49] osm: we dont suport mysql [20:52] osm, in a bit more detail, we use triggers and other pgsql-specific functionality [20:52] osm, do you have any more context to your request? why mysql? [21:01] roaksoax/kiko: If I "auto assign" an interface will the node I deployed "always" have that IP? [21:02] for example if I release the node X and then image 4 other nodes on the same network (with auto assign) will they ever get the IP that node X had? [21:02] shewless, I am pretty sure they may [21:03] shewless, i.e. we will need to recycle IPs at some point if the machines are released, right? [21:03] kiko: right.. just thinking maybe I'll do "static" [21:04] I have a separate DNS server and I want it to be deterministic [21:04] roaksoak/kiko: I have some project with mysql backend and i want to integrate maas. it should be possible with django [21:04] osm, if only it was so simple. we really are tied to pgsql :) [21:13] kiko: how much tied? I tried to change it and get error maasserver.Domain.name: (mysql.E001) MySQL does not allow unique CharFields to have a max_length > 255. [21:13] is it only model or methods as well? [21:16] osm, I suggest deploying MAAS with pgsql and then evaluating the DB schema [21:16] it's a pretty deep dependency [21:16] but you can just hide pgsql on the same node MAAS runs [21:16] and consume the API as if it were, say, sqllite [21:18] yes, it could be solution. Thanks for your advise [21:18] thanks for the questions [21:35] shewless: the "auto assign" IP address may be recycled and given to another machine at any given point, if you "static assign" it will always used for that one [21:36] shewless: i'd recommend you use static assign if you already degining those in your external DNS server [23:19] roaksoax/kiko:where can I get support for MAAS?