* [PVE-User] Bond and hardware parameters...
@ 2021-03-27 18:08 Marco Gaiarin
0 siblings, 0 replies; only message in thread
From: Marco Gaiarin @ 2021-03-27 18:08 UTC (permalink / raw)
To: pve-user
I don't know if this can be considered a 'bug', or a 'feature request',
for now i make a note and hope in some feedback.
I need in a PVE server to have a bond interface but i need also to set
some hardware parameters, following the 'bufferbloat' principle, eg no
performance, full control. This bond have to manage some (slow) WAN
link, that have to be policed.
I've tried to setup some parameters, eg installed 'ethtool' and added
to slave bond interfaces:
iface enp48s0 inet manual
ethernet-autoneg on
link-speed 100
link-duplex full
hardware-dma-ring-tx 64
offload-tso off
offload-gso off
offload-gro off
iface enp48s1 inet manual
ethernet-autoneg on
link-speed 100
link-duplex full
hardware-dma-ring-tx 64
offload-tso off
offload-gso off
offload-gro off
auto bond0
iface bond0 inet manual
slaves enp48s0 enp48s1
bond_miimon 100
bond_mode 802.3ad
bond_xmit_hash_policy layer2+3
bond-slaves enp48s0 enp48s1
bond-downdelay 200
bond-updelay 200
but found that parameters was not set. But if i fire up the interface
and then the bond by hand using ifenslave, the hardware parametrs seems
to be kept.
After fiddling a bit, i've found in /usr/share/doc/ifenslave/README.Debian.gz:
Using ifup on a master interface will call ifup for all slaves that are
flagged with allow-bondX. (bondX being replaced by the master interface
name). This will allow for extra setup for special slave interfaces.
so seems that the correct syntiax is:
iface enp48s0 inet manual
ethernet-autoneg on
link-speed 100
link-duplex full
hardware-dma-ring-tx 64
offload-tso off
offload-gso off
offload-gro off
allow-bond0
iface enp48s1 inet manual
ethernet-autoneg on
link-speed 100
link-duplex full
hardware-dma-ring-tx 64
offload-tso off
offload-gso off
offload-gro off
allow-bond0
auto bond0
iface bond0 inet manual
slaves enp48s0 enp48s1
bond_miimon 100
bond_mode 802.3ad
bond_xmit_hash_policy layer2+3
bond-slaves enp48s0 enp48s1
bond-downdelay 200
bond-updelay 200
I've tried in this way, and effectively hardware parameters are set
correctly, as expected.
Indeed, probably this will hit also MTU settings, but probably the MTU
setting in an LACP bond 'propagate' automaticaly to slaves, so i've
never noted that.
Say me if it is better to fire up a bug, to have PVE web interface to
automatically accomodate that. Thanks.
--
dott. Marco Gaiarin GNUPG Key ID: 240A3D66
Associazione ``La Nostra Famiglia'' http://www.lanostrafamiglia.it/
Polo FVG - Via della Bontà, 7 - 33078 - San Vito al Tagliamento (PN)
marco.gaiarin(at)lanostrafamiglia.it t +39-0434-842711 f +39-0434-842797
Dona il 5 PER MILLE a LA NOSTRA FAMIGLIA!
http://www.lanostrafamiglia.it/index.php/it/sostienici/5x1000
(cf 00307430132, categoria ONLUS oppure RICERCA SANITARIA)
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2021-03-27 18:10 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-03-27 18:08 [PVE-User] Bond and hardware parameters Marco Gaiarin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox