[ovirt-users] bonding interfaces

[Bill Dossett]

Hi, no network manager... I have been sort of doing what you describe.  I create my first interface manually with a minimum configuration (when I was first testing ovirt, I was using trunked ports and vlans, so had to do all that manually) then add the host to ovirt-engine and it did it's magic and bonded interfaces for me.. I'm going to try a later release I think as it seems there may have been a bug like this in 3.5.1.

Thank you

From: Soeren Malchow [mailto:soeren.malchow at mcon.net]
Sent: 19 April 2015 14:48
To: Bill Dossett; users at ovirt.org
Subject: Re: [ovirt-users] bonding interfaces

Hi Bill,

First, i would strongly suggest that you check whether the interfaces are explcicitly NOT managed by NetworkManager, thats the way we do it.

We have something like this

-

DEVICE=ovirtmgmt
TYPE=Bridge
DELAY=0
STP=off
ONBOOT=yes
IPADDR=
NETMASK=
GATEWAY=
BOOTPROTO=none
DEFROUTE=yes
NM_CONTROLLED=no
HOTPLUG=no

----

And

-
DEVICE=bond0
BONDING_OPTS='miimon=100 updelay=0 downdelay=0 mode=802.3ad xmit_hash_policy=layer3+4'
BRIDGE=ovirtmgmt
ONBOOT=yes
NM_CONTROLLED=no
HOTPLUG=no
---

Physical interfaces

-

DEVICE=p3p3
HWADDR=
MASTER=bond0
SLAVE=yes
ONBOOT=yes
MTU=1500
NM_CONTROLLED=no

-

And we actually create all the interfaces BEFORE ovirt installation even though the files will be regenerated after installing and configuring ovirt (vdsm)

That works like a charm for us

Cheers
Soeren


From: Bill Dossett <Bill.Dossett at pb.com<mailto:Bill.Dossett at pb.com>>
Date: Thursday 16 April 2015 21:49
To: "users at ovirt.org<mailto:users at ovirt.org>" <users at ovirt.org<mailto:users at ovirt.org>>
Subject: [ovirt-users] bonding interfaces

Hi,

I had a Gluster Cluster running that I setup with oVirt engine... "had" being the operative word here.

Today I connected the second network interface and then went to the host and tried to bond this interface to the ovirt mgmt. network... this is where the "had" part comes in ;-)

Now I have no working interfaces at all, can't ping in or out and all interfaces say they are down, en1, 2, ovrtmgmt and bond0 when I do an ip add (this is centos7)

I have done this quite a few times on RHEV clusters and it always worked well, very impressed with it.

The differences are...

RHEV cluster I was using trunked ports and VLANS
RHEV I was using 6.4 or 6.5 RHEL

I think that's it.

I haven't really dug into the /etc/sysconfig/network-scripts to see what normally goes into the config files to create a bonded interface - only to see how to configure my first interface with a vlan and then I let ovirtengine do the rest for me as it worked so wellI

I guess I may have to rip it out and start over unless anyone has any tips on fixing it.  It's just PoC anyway, so learning.

Thanks


Bill Dossett
Systems Architect

Tech Central - Global Engineering Services

T +1 303 440 3523
M +44 (0)777 590 8612
bill.dossett at pb.com<mailto:bill.dossett at pb.com>
pitneybowes.com<http://pitneybowes.com/>

Pitney Bowes
4750 Walnut Street | Boulder, Colorado, 80301 | USA




In Engineering?
Raise a ticket via Remedy Anywhere [HERE<applewebdata://5D25CA10-0BA9-4233-9DDF-69951F87F8AB/teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident/newIncident.aspx>] takes less than a minute

CloudForms User Guide available [HERE<https://dl.dropboxusercontent.com/u/6219441/ManageIQ%20-%20User%20Manual%20PB%20v5.pdf>]




________________________________


________________________________

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20150420/4438a703/attachment-0001.html>