1a) Correct. At the time we first did it, iPXE codebase and/or uefi didn't do well at 'exiting', so offering nothing was the choice, while pxe is slower, it means we didn't have to contend with UEFI crashes that happened in the wake of iPXE trying to exit. Things might have changed by now between UEFI maturing and iPXE maturing, but even with kkpxe in legacy we occasionally have BIOSes where exit to short out the netboot attempt causes headaches. Considered a little less critical as UEFI OSes tend to rewrite the bootorder on deploy to make themselves first, and we have the ability to explicitly request network boot through a BMC by and large, so the fastest mechanism for diskful boot is to just boot straight to hard drive and only network boot when there's a deployment to do. Incidentally, some security teams have started requiring this, as they don't like the concept of a PXE attempt every boot.
-----Original Message----- From: THomas HUMMEL <thomas.hum...@pasteur.fr> Sent: Monday, October 11, 2021 9:37 AM To: xcat-user@lists.sourceforge.net Subject: [External] Re: [xcat-user] xCAT 2.16.2 new xNBA issue Hello Mark, So if I sum up again, what was left unclear to me was: a) is it expected that after the initial install, when a UEFI stateful node reboots in PXE first (if order gets manually changed again for instance), no tftp of xNBA occurs ? : like xCAT would check the install status of the node ? I've always thought the way stateful once installed booted on disk (if PXE first) was to download the dummy 'exit' only iPXE script via xNBA (which would imply to see tftp of xNBA beforehand), not by giving up on no boot filename answer b) why isn't the .uefi /tftpboot file (/tftpboot/xcat/xnba/nodes/<node>.uefi) changed in sync with the no-extension one after install ? c) for duplicates messages : I still don't know why those messages show in logs. Thanks for your help -- Thomas HUMMEL On 10/1/21 18:19, THomas HUMMEL wrote: > [Sorry Mark for the duplicate answer - I mistakenly reply to you only] > > On 10/1/21 17:26, Mark Gurevich wrote: >> Thomas, >> >> Sorry, I do not quite follow your point about "2)", is the observed >> behavior different from what you expected ? > > Well yes : I would expect (expect may not be the correct word - let's > say I thought it worked that way) that a node, even once installed (in > the stateful case), which reboots on the network would tftp xNBA > (which in turn would GET one of the /tftpboot/xcat/xnba/node/<node> or > <node>.uefi script). > > My understanding is that it works this way for legacy (BIOS) boot. > Otherwise what would be the point to change the iPXE script file to: > > #!gpxe > #boot > exit > > ? > > So I was assuming any host PXE-ing would always get xNBA wether it is > on initial install or a later reboot (if UEFI network is first in boot > order) > > >> >> As to "duplicate lease" point, is it possible you have an IP >> associated with a MAC, that is also inside dynamic range ? > > I thought about it but I don't thinks it's the case. > >> Can you show: >> >> lsdef -t network -l > > > "maestro_net","192.168.144.0","255.255.240.0","eth0","192.168.144.1",, > "<xcatmaster>",,,,"192.168.144.2-192.168.147.254",,,,,,"maestro.pasteu > r.fr","1500",, > > "maestro_ipmi","10.7.96.0","255.255.248.0",,"10.7.96.1",,,,,,,,,,,,,"1 > 500",, > > "maestro_ipoib","172.16.0.0","255.255.248.0",,,,,,,,,,,,,,,"1500",, > >> lsdef <node> -i ip,mac > > Object name: maestro-satvmtmpl > ip=192.168.148.204 > mac=00:50:56:b5:ac:de > >> makedhcp -q node > > # makedhcp -q maestro-satvmtmpl > maestro-satvmtmpl: ip-address = 192.168.148.204, hardware-address = > 00:50:56:b5:ac:de > > Thanks for your help > _______________________________________________ xCAT-user mailing list xCAT-user@lists.sourceforge.net https://apc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fxcat-user&data=04%7C01%7Cjjohnson2%40lenovo.com%7Cf12a7ae143734d28f70108d98cbc4565%7C5c7d0b28bdf8410caa934df372b16203%7C1%7C0%7C637695562527776411%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=1uZ2Ak81i5pksIjigqV0Gg1VOkBskHdpT2a3EBRsoaA%3D&reserved=0 _______________________________________________ xCAT-user mailing list xCAT-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xcat-user
