[12:19] <PTO> Im trying to use MAAS on an IBM Bladecenter H, and when a node enlist in MAAS, the power IPMI configuration gets an internal IP 192.199.199.81 of the blade. How do I get around this in MAAS? Should I specify the AMM module IP instead or try to route the chasis internal network on the LAN?
[12:44] <parlos> @PTO I'm about to embark on a similar path; (same bladecenter). Does the MAAS correctly detect the blade?
[13:27] <roaksoax> PTO: maas can only manage machines inside the blade, not the blade itself
[13:27] <roaksoax> parlos: ^^
[13:30] <parlos> roaksoax, hrm, so the normal way to commision will be problematic (as that needs power control).
[13:33] <roaksoax> parlos: right, well maas will need to communicate directly to the machines. That said, I thought that was possible in these blades ? is it not ?
[13:33] <parlos> roaksoax I hope so..
[13:33] <parlos> Have to do a bit of RFTM then :)
[13:38] <parlos> I was quietly hoping that the MAAS 'detection' image (ie. the PXE booted kernel) would be smart enough to detect the BMC on the blade when it booted..
[13:44] <roaksoax> parlos: well there's no way for the machine to detect the blade, unless maas specifically supports the balde and you tell maas that there's a blade (in which case maas would discover the machines inside the blade)
[13:47] <parlos> roaksoax; the terminology Im used to is that a (IBM flavoured i guess), is: BladeCenter hosts blades, the blades are complete computers (cpu,mem and disc).
[13:49] <parlos> The major difference between a blade and a regular computer is the I/O connectivity.
[13:50] <parlos> However, if i compare to a 'real' device, im not sure where the BMC would be located. As, it would seem that it would be a part of the BladeCenter, not the blade it self (Cf. powering VMs)
[13:50] <parlos> However, if i compare to a 'real' device, im not sure where the BMC would be located. As, it would seem that it would be a part of the BladeCenter, not the blade it self (Cf. powering VMs via SSH)
[13:51] <PTO> roaksoax: When a blade PXE boots, it goes throught the rutines and powers off. It will enlist in mass and the power configuration is IPMI2 with the IP address 192.199.199.81 (blade1). To my understanding the Chassis internal Network (cin) is not accessible outside the bladecenter. Google'ing suggests that you configure the bladecenter AMM module as the IPMI host, but i cant quite figure out if this is the way to go?
[14:17] <parlos> PTO Is there an completely internal network? In my case we have two switches in the back (one active, one active-standby). But they are 'directly' connected to the blades.
[14:27] <roaksoax> so MAAS can manage machines in two ways, via a chassis or directly
[14:27] <roaksoax> via a chassis means that maas is told about a chassis and it discovers the machines inside, and power manages them via the chassis
[14:28] <roaksoax> directly, when machines have BMC's and maas can power managed them directly
[14:28] <roaksoax> we dont support bladecenter H as a "chassis" in MAAS, so to power manage the machine, we need to manage them directly
[14:28] <roaksoax> my understanding is that this is possible
[14:41] <PTO> roaksoax: How do you instruct MAAS to use the chais and not the IMM of the server?
[14:43] <roaksoax> PTO: you would need to create a new power driver
[14:56] <PTO> Bump... Thats not what i want to hear :-/
[17:01] <Charlie> Hi, would any one be able to offer some help? Just set up my first MAAS environment, added a physical node on the network MAAS looks at with DHCP/DNS. initially all looks to be ok and then it fails to commission my node.
[17:06] <Charlie> being new the MAAS, im having trouble understanding why this is failing
[17:07] <Charlie> looking in /var/log/maas/rsyslog/ for the node, I can see cloud-final.service: Main process exited, code=exited, status=1/FAILURE
[17:07] <Charlie> google tells me very little about this.
[17:11] <Charlie> Would anyone be able to assist?
[23:30] <edmz> how can I find the reason of a "Node installation failure - 'curtin' curtin command install" failed deployment?
[23:30] <edmz> this node was deployed. I decided to re-deploy and got this error.