[lustre-discuss] lnet ip aliases

[Horn, Chris]

Yeah, I think LNet needs the address to be tagged with a unique label. The patch https://review.whamcloud.com/c/fs/lustre-release/+/53605 may allow you to configure the desired NIDs without the labels.

Chris Horn

From: lustre-discuss <lustre-discuss-bounces at lists.lustre.org> on behalf of Michael DiDomenico via lustre-discuss <lustre-discuss at lists.lustre.org>
Date: Tuesday, April 23, 2024 at 8:54 PM
To: lustre-discuss <lustre-discuss at lists.lustre.org>
Subject: [lustre-discuss] lnet ip aliases
in the past under rhel7 i would do this

[root]# ip add add 192.168.12.29/24<http://192.168.12.29/24> dev enp0s13f0u1 label enp0s13f0u1:1 broadcast 192.168.12.255 noprefixroute
[root]# ip add show enp0s13f0u1
2: enp0s13f0u1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether a0:ce:c8:b2:22:f7 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.29/24<http://192.168.10.29/24> brd 192.168.10.255 scope global noprefixroute enp0s13f0u1
       valid_lft forever preferred_lft forever
    inet 192.168.12.29/24<http://192.168.12.29/24> brd 192.168.12.255 scope global noprefixroute enp0s13f0u1:1
       valid_lft forever preferred_lft forever

[root at cyan ~]# lnetctl net add --ip2net="tcp1(enp0s13f0u1) 192.168.10.*"
[root at cyan ~]# lnetctl net add --ip2net="tcp2(enp0s13f0u1:1) 192.168.12.*"
[root at cyan ~]# lctl list_nids
192.168.10.29 at tcp1
192.168.12.29 at tcp2

but i've upgraded to redhat9 using network manager fully now.  when i do

[root at cyan system-connections]# nmcli c modify Wired\ connection\ 1 ipv4.addresses "192.168.10.29/24,192.168.12.29/24<http://192.168.10.29/24,192.168.12.29/24>"
[root at cyan system-connections]# nmcli c up Wired\ connection\ 1
Connection successfully activated (D-Bus active path: /org/freedesktop/NetworkManager/ActiveConnection/22)
[root at cyan system-connections]# ip add show enp0s13f0u1
2: enp0s13f0u1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether a0:ce:c8:b2:22:f7 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.29/24<http://192.168.10.29/24> brd 192.168.10.255 scope global noprefixroute enp0s13f0u1
       valid_lft forever preferred_lft forever
    inet 192.168.12.29/24<http://192.168.12.29/24> brd 192.168.12.255 scope global noprefixroute enp0s13f0u1
       valid_lft forever preferred_lft forever

this doesn't work

[root at cyan ~]# lnetctl net add --ip2net="tcp1(enp0s13f0u1) 192.168.10.*"
[root at cyan ~]# lnetctl net add --ip2net="tcp2(enp0s13f0u1) 192.168.12.*"
[root at cyan ~]# lctl list_nids
192.168.10.29 at tcp1
192.168.10.29 at tcp2

i'm pretty sure this is expected behavior based on how lnet works, but i figured i'd ask around


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20240425/0d4016db/attachment.htm>