[00:43] I upgraded 1.7.5 to 1.8.0 and I'm getting a 403 on /MAAS now :( [00:44] here are /var/log/maas/maas.log http://paste.ubuntu.com/11807587/ [00:45] apache2/error.log http://paste.ubuntu.com/11807588/ [00:46] apache2/access.log http://paste.ubuntu.com/11807589/ [00:48] traceback in /var/log/maas/regiond.log http://paste.ubuntu.com/11807608/ [01:02] I reinstalled MAAS, I'm able to access the web UI, but I'm getting this message "One cluster is not yet connected to the region. Visit the clusters page for more information." [01:02] I can't seem to get it connected... [01:08] Doing what I've done, all my nodes are associated with an old cluster -- how can I change that association? [01:12] I edited /etc/maas/maas_cluster.conf with the old UUID -- I'm back in business I think! [02:01] In 1.8.0 how do I define what OS gets installed? Before I could acquire, then edit the node, and choose the OS. Here, the OS is not a choice when I edit the node... [02:04] Ah..there it is, when I choose deploy, I can choose the OS! [02:15] Is there a way to get a kernel newer than hwe-t on MAAS? I'm looking for 3.19 kernel... [02:23] Figured that one out as well. Guess I need the whole image via the 'images' tab -- now I have hwe-u/v :D [09:34] rvba: by the way, I'm adding devices support to gomaasapi [09:34] rvba: WIP mp here if you wanted to take a look [09:34] https://code.launchpad.net/~mfoord/gomaasapi/devices/+merge/263370 [09:34] rvba: I expect it to be finished today [09:34] voidspace: cool… not sure I'll be able to have a look today but I'll try… I'm adding it to my list for tomorrow morning :) [09:34] rvba: we're only implementing the subset (large) of devices API that we use, with TODOs marking the bits that haven't been done if people want to add them [09:35] rvba: cool, thanks === frankban_ is now known as frankban === jfarschman is now known as MilesDenver === liam_ is now known as Guest41856 [16:12] rvba: ping [16:13] as far as I can tell, using the devices api claim-sticky-ip-address does nothing (or at least doesn't do what I would expect) [16:14] if I call it on a node, I get success result but there's nothing in ip_addreses for the device [16:14] and it's resolutely empty after reading too [16:16] it's not actually a problem for us, we're specifying IP address [16:16] it's just surprising [16:17] I'll file an issue [16:17] Morning folks -- I've asked before, but I hope for a clearer answer. Is there a way to have local password for console access? In the event that I can't SSH into a newly deployed server. [16:21] bug 1470930 [16:31] If I want to create a custom curtin installer for an hwe-t kernel (rather than generic), would I do custom_ubuntu_amd64_hwe-t or hwet or ? [17:50] What's the equivalent of this in curtin? d-i passwd/root-password-crypted password $HASH [17:53] bmorriso, I would guess it would require being added as curtin userdata [17:53] bmorriso, you want that run at the end of the installation? [17:55] I'm not sure where it needs to be ran/added. I guess I could just do it in a late command setting the password for the root user (which should enable the root user) [17:55] I worked around the lack of a local ubuntu access by just creating a user account in a late command. [18:38] somebody kowns how to set the power type in maas, my node is running on vmware fustion 7.1.1, and the node was added to mass "Nodes" list. === CyberJacob is now known as zz_CyberJacob [18:40] my maas version is 1.8 [18:40] hezhiqiang, edit the node, basically? [18:40] Yes, i entered the edit page [18:41] But i down konw how to fill in those fileds [18:42] These fileds are: Power type, VM Name (if UUID unknown), VM UUID (if known), VMware hostname, VMware username, VMware password, VMware API port (optional), VMware API protocol (optional) [18:43] In Power type droplist, I selected the "Vmware" item [18:46] mpontillo, ^^^ [18:47] hezhiqiang: unfortunately, VMware Fusion does not support the same remote-access API as Workstation, ESXi, and vSphere, so is not currently supported [18:47] * kiko . o O ( what is vmware fusion )? [18:47] -_-! [18:48] hezhiqiang: however, I have successfully tested ESXi running as a guest under VMware Fusion ;-) [18:48] ah, it's a vmware for macos thing [18:49] mpontillo, is fusion basically the brand for vmware on macos? or is it a mac-specific product? [18:49] kiko: it's basically VMware Workstation, but optimized for the Mac, and doesn't support some of the advanced features of VMware Fusion [18:49] kiko: rather, the advanced features of VMware Workstation [18:50] kiko: it's Mac-specific (and the only product they support on the Mac) [18:50] including the API [18:50] that is correct kiko [18:50] mpontillo, some background on vmware -- workstation and esxi are single-node right? [18:50] whereas vsphere lets you manage VMs on multiple nodes? [18:51] This is the error message: Failed to power on node - Node could not be powered on: vmware failed with return code 1: [Errno 2] No such file or directory: '://few-use://sdk/vimServiceVersions.xml' [18:53] It's supported on Vmware workstation on Linux ? [18:53] kiko: not sure what you mean by "single-node"; if you mean a single hypervisor, yes. vSphere allows you to do things like clustering, failover, vMotion, etc [18:53] hezhiqiang: yes, I test MAAS with VMware Workstation frequently, so I know that works. [18:53] mpontillo, well, meaning they are supposed to run on one physical machine. [18:54] kiko: right, as opposed to multiple (vSphere) [18:54] so I think yes, a single hypervisor [18:54] mpontillo: thanks, i will try it on my linux box [18:54] mpontillo, what does vsphere call what runs on each of the nodes it controls? [18:54] hezhiqiang: no problem; please let me know how it works [18:55] kiko: I'm not sure what terminology they use, but it's basically multiple instances of the ESXi hypervisor which are managed by vSphere [18:55] you can download ESXi for free and run it on physical hardware (or a supported hypervisor, like vmware fusion or workstation), and MAAS can manage the nodes on a bare metal ESXi box as well [18:58] hezhiqiang: one thing to note about VMware Workstation, is that you need to put the VMs you want MAAS to manage under the "Shared VMs" section. [18:58] after you do that, the API can see them [19:00] hezhiqiang: also, I recommend you use the "Add Chassis" button to discover the VMs. that way you won't need to worry about setting the power parameters, BIOS boot order, etc [19:01] just note that when you do this, it will change the boot order to (pxe, hard disk) for each VM that matches the prefix filter === zz_CyberJacob is now known as CyberJacob [19:02] mpontillo: I kown, i was readed about MAAS 1.8 changelog, It requires the python-pyvmomi package to be installed [19:02] right. [19:03] I'll go to bed, there is +8:00 timezone, have a nice day, guys. [19:03] good night hezhiqiang [19:23] I'm curious, can I target tags in curtin? [19:31] bmorriso, hmm, what do you mean? [19:31] if I tag a host "dev" or "staging" or "prod" could I have late_commands that run conditionally based on matching tag? [19:31] If dev == run this command elif staging == run this command, else == run this command [19:33] I don't want to have a curtin installer per host, so I'm hoping I can do it at a more generic level with some logic in the template [19:34] Like I see this example http://astokes.org/customizing-fastpath-curtin-installations/ and I don't know if "http_proxy" is a tag or where it is coming from [19:54] ah [19:54] smoser, are node tags visible from curtin userdata? [19:54] bmorriso, node.tags perhaps? [19:55] ah [19:55] bmorriso, http://askubuntu.com/questions/466469/maas-use-different-preseeds-scripts-based-on-zone-or-tag [20:06] i was using a config file stored in the static folder of MAAS webserver. however 'tags' is a much cleaner and generic approach. thanks bmorriso, kiko :) [20:20] yw! [21:21] I'll give it a shot! Thanks!! === CyberJacob is now known as zz_CyberJacob === zz_CyberJacob is now known as CyberJacob [22:20] Any idea why IPMI would sort of just stop working? "Jul 2 15:19:18 maaspoc maas.websockets.node: [ERROR] maas-test-box: Timed out waiting for power response in Node.power_state" [22:20] This host was working just fine earlier, now it is getting into weird states with IPMI/Power [22:24] my maas server's ip address changed - i dpkg-reconfigured everything to use the new ip and it works until curtin, but curtin tries to download the tarball from the old ip [22:24] how do i fix that up? [22:27] what does /etc/maas/maas_cluster.conf look like? [22:28] or /etc/maas/maas_local_settings.py? [22:28] Anyone ever seen this before? http://i.imgur.com/4t6HskW.png kiko? [22:28] bmorriso: both have the new ip [22:32] Do you have the correct settings under the network tab? === CyberJacob is now known as zz_CyberJacob [22:42] bmorriso: yeah [22:42] bmorriso: only thing that changed was my ip - subnet is the same, etc. only place that ip appears to be is in the postgres db [22:42] s/that ip/the old ip/ === zz_CyberJacob is now known as CyberJacob [22:44] restarted apache and maas-clusterd? [22:49] bmorriso: even rebooted :) [22:49] wow [22:50] I've changed my IP before, but I don't remember it being that difficult. What version? [22:50] bmorriso: 1.8 [22:51] When I changed my IP, it was 1.6. 1.8 is definitely a bit different :-/ === CyberJacob is now known as zz_CyberJacob