| *** rcw has quit IRC | 00:25 | |
| *** sgw has quit IRC | 05:03 | |
| *** hyunsikyang__ has joined #starlingx | 06:48 | |
| *** hyunsikyang has quit IRC | 06:52 | |
| *** bengates has joined #starlingx | 07:19 | |
| *** TheJulia has quit IRC | 07:56 | |
| *** mnasiadka has quit IRC | 07:56 | |
| *** TheJulia has joined #starlingx | 08:01 | |
| *** mnasiadka has joined #starlingx | 08:01 | |
| *** irclogbot_3 has quit IRC | 09:41 | |
| *** irclogbot_2 has joined #starlingx | 09:43 | |
| *** hyunsikyang has joined #starlingx | 12:54 | |
| *** hyunsikyang__ has quit IRC | 12:58 | |
| *** billzvonar has joined #starlingx | 13:19 | |
| *** sgw has joined #starlingx | 13:43 | |
| *** stampeder has joined #starlingx | 13:56 | |
| stampeder | Good morning. From the documentation on Duplex mode it appears that one needs to configure a PXE boot server in order to first boot Controller-1? Is this correct? | 13:59 |
|---|---|---|
| bwensley | Controller-1 will PXE boot from controller-0. The PXE boot server is automatically configured on controller-0. | 14:01 |
| stampeder | bwensley: Thanks. So one would just point the PXE boot ip to controller-0's IP? | 14:19 |
| stampeder | In the bios I mean. | 14:19 |
| bwensley | You need to force controller-1 to network boot from controller-0. | 14:26 |
| bwensley | The boot order should be set to disk first, then network. | 14:26 |
| bwensley | For the initial install of controller-1, you either need to wipe the disk (so it won't boot from disk) or force the boot from the network. | 14:26 |
| stampeder | I'm using a Dell C6100 with four blades. I have controller-0 on blade 1. Controller-1 is on blade 3 for now and both connect into a layer 2 switch. Controller-1 network boots ( I can set the boot priority to network to force this) and goes looking for DHCP but it never seems to find controller-0 as dhcp. Is there a way I can confirm that | 14:38 |
| stampeder | controller-0 is capable of responding to the DHCP request? | 14:38 |
| mfe555 | Hi. Reading the documentation I see a limit of 100 worker nodes. Why 100? | 14:47 |
| sgw | stampeder: correct, the active controller has dhcp and pxe on the "mgnt" | 14:57 |
| sgw | network | 14:57 |
| stampeder | The simplex setup does not implement the "mgnt" network. My servers have two data ethernet ports. They also have a third BMC port that can be shared although the bios does not see it. It can be configured as IPMI. I setup OAM to use the first data port along with this port functtioniing as my main data port. Should I have setup data port 1 as my | 15:02 |
| stampeder | external data port and data port 2 as the OAM port? Thanks. | 15:02 |
| stampeder | In a normal data center we would setup ports 1 and 2 as data ports capable of LAG. The BMC would be our out of band management port that responds to IPMI. Sometimes a four ethernet port is added as a MGNT (inband) lan port. | 15:04 |
| *** bengates has quit IRC | 16:45 | |
| stampeder | Is there a StarlingX command cheat sheet available? | 17:25 |
| stampeder | bwensley: controller-0 running snd unlocked. Controller-1 has empty disk (formatted ext4 only). Both plugged into adjacent ports on layer 2 switch. Controller-1 force to PXE boot on network. Controller-1 console posts PXE-E51: NoDHCP or proxyDHCP offers were received...Any thoughts? | 17:39 |
| bwensley | I'm not a networking guy, but I expect something in your network config is preventing the DHCP requests from getting from controller-1 to controller-0. | 17:42 |
| stampeder | dwensley: I did a netstat -an | fgrep -w 67 to verify that controller-0 was listening on port 67 and it came back with the correct status. So, it appears to be listening but not offering. The only place that a firewall is involved would be in controller-0 as it uses ip-tables. Thoughts? | 17:54 |
| bwensley | stampeder: we have done thousands of controller-1 installs over the years. I'm 99.9% sure that controller-0 is configured properly to reply to the DHCP requests. | 18:14 |
| bwensley | Unless you have changed something on controller-0 (have you added iptables rules?), this has got to be a configuration/switch/networking issue. | 18:15 |
| sgw | slittle1: patch posted for puppet-modules https://review.opendev.org/726219 Bumped TIS_PATCH_VER | 18:32 |
| stampeder | bwensley: I have done zero controller-1 installs so I am quite in the crawling stage at this. From what I have seen it uses DNSMASQ with which I am fairly familiar. I'm also using Intel Boot Agent GE v1.3.64 which should be working very well as it seems to be sending out DHCP requests. This leaves me at the point of breaking out Wireshark and start | 19:07 |
| stampeder | looking at packets. Thanks for helping me narrow things down. I very much appreciate it. | 19:07 |
| *** yaawang has quit IRC | 19:16 | |
| bwensley | Good luck - let us know what you find. | 19:40 |
| *** yaawang has joined #starlingx | 20:32 | |
| *** hoonetorg has quit IRC | 22:34 | |
| *** hoonetorg has joined #starlingx | 22:46 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!