[ale] Getting rid of VMware
Derek Atkins
derek at ihtfp.com
Fri Mar 12 10:47:38 EST 2021
Linux absolutely has LAG support (as does FreeNAS, by the way, which you
might want to consider if you're building a dedicated file server).
I'm going to quote RH/Fedora configs, but... Basically you need to create
a bond0 interface:
DEVICE=bond0
BONDING_OPTS='mode=4 lacp_rate=1'
ONBOOT=yes
MTU=1500
DEFROUTE=no
NM_CONTROLLED=no
IPV6INIT=no
[ and then your IP configs ]
and then, assuming you have e.g. eno1 and eno2:
DEVICE=eno1
MASTER=bond0
SLAVE=yes
ONBOOT=yes
MTU=1500
DEFROUTE=no
NM_CONTROLLED=no
IPV6INIT=no
DEVICE=eno2
MASTER=bond0
SLAVE=yes
ONBOOT=yes
MTU=1500
DEFROUTE=no
NM_CONTROLLED=no
IPV6INIT=no
The "mode=4" means to use LACP. There are other kinds of bond modes.
>From "modinfo bonding":
parm: mode:Mode of operation; 0 for balance-rr, 1 for
active-backup, 2 for balance-xor, 3 for broadcast, 4 for 802.3ad, 5 for
balance-tlb, 6 for balance-alb (charp)
Hope this helps.
-derek
On Fri, March 12, 2021 10:28 am, Tod Fassl wrote:
> I created a Win10 VirtualBox virtual machine on my desktop yesterday. I
> got the VMWare client thingy installed and working. I'll start poking
> around in there today.
>
>
> Thanks for the tip on link aggregation group. That's exactly the kind of
> tip I needed just so I know what to google. Here's a really good
> explanation of LAG:
>
> https://cloud-duo.com/2019/05/lacp-configuration-with-vmware-esxi-part-1/
>
>
> So I am thinking we may see a performance degradation unless I can find
> the Linux equivalent of LAG.
>
>
> On 3/12/21 8:15 AM, Derek Atkins wrote:
>> Hi,
>>
>> My guess is that the multiple ethernet ports are being used as a Link
>> Aggregation Group (which would allow you to have a single IP address but
>> increase the aggregated network capabilities). But you'd have to check.
>>
>> I think you will need to at least look at the VMware host configuration
>> to
>> figure out how the host *is* configured. I don't think you need to be a
>> vmware expert, but you will need to login as vmware admin to look at the
>> configs.
>>
>> -derek
>>
>> On Fri, March 12, 2021 9:08 am, Tod Fassl via Ale wrote:
>>> We had four machines in the cluster. We've already pulled 3 of them
>>> out.
>>> Two, I've already installed Linux on and have been repurposed. One is
>>> sitting there waiting for me to wipe out ESXI and install Linux on it.
>>> Our virtual Linux file server is still running on that one remaining
>>> ESXI host. So my task is to install Linux on that machine we just
>>> pulled
>>> out of the cluster and reconnect the ISCSI array to it.
>>>
>>>
>>> What I need to get a grip on is basically how it is working now. I
>>> don't
>>> want to have to become an VMWare expert for a week of work. I suspect
>>> that what I will be able to do is to follow the usual steps for
>>> attaching an ISCSI array to a Linux file server. But I don't understand
>>> why there are all these cables and how its working now.
>>>
>>>
>>> The virtual machine that is acting as a file serveris running on a ESXI
>>> host that has 6 ethernet cables connected to it. But it looks like most
>>> of the ports aren't even active. I would *assume* I can safely remove
>>> those cables. But why the heck are they there in the first place?
>>>
>>>
>>> There's this famous anecdote in comp sci. There was a bare copper wire
>>> running from an old room-sized computer to a filing cabinet. There was
>>> a
>>> note on the filing cabinet that said, "Do not remove this cable." So a
>>> new guy comes in, first thing he does is remove the cable, and of
>>> course, the machine crashes.
>>>
>>>
>>> I don't want to be that guy.
>>>
>>>
>>> I guess I'll have to bite the bullet and learn more about VMWare. Kinda
>>> sucks though. I'm a math major, not a brick layer.
>>>
>>>
>>> On 3/11/21 5:13 PM, Jim Kinney wrote:
>>>> That sounds like the hardware nodes supports the iscsi protocol or the
>>>> base hypervisor layer from vmware is doing the iscsi connection.
>>>> Probably the later.
>>>>
>>>> Either way, you're going have a challenge unless you can get details
>>>> on the iscsi contents. You'll need to tie each virtual drive to it's
>>>> correct device. Otherwise virtual machine A gets the drive space of
>>>> virtual machine B.
>>>>
>>>> From ESX you can get the UUID string of the virtual drive used by
>>>> that
>>>> linux vm.
>>>>
>>>> I would use that vm now as just a source for a backup. Then restore to
>>>> the new hardware machine running a base install with restore bits. The
>>>> iscsi array will still need a partition to use for the hardware linux.
>>>> I doubt vmware will make this easy.
>>>>
>>>> Un-virtualizing the drive on an iscsi array sounds like lots of pain.
>>>> Retire first.
>>>>
>>>> On March 11, 2021 5:37:36 PM EST, Tod Fassl via Ale <ale at ale.org>
>>>> wrote:
>>>>
>>>> Soonish, I am going to have to take an ISCSI array that is
>>>> currently
>>>> talking to a VMWare virtual machine running Linux and connect it
>>>> to
>>>> a
>>>> real Linux machine. The problem is that I don't know how the
>>>> Linux
>>>> virtual machine talks to the array. It appears as /dev/sdb on the
>>>> Linux
>>>> virtual machine and is mounted via /etc/fstab like its just a
>>>> regular HD
>>>> on the machine.
>>>>
>>>>
>>>> So I figure some explanation of how we got here is in order. My
>>>> previous
>>>> boss bought VMWare thinking we could take 4 24-core machines and
>>>> make
>>>> one big 96-core virtual machine out of them. He has since
>>>> retired.
>>>> Since
>>>> I was rather skeptical of VMWare from the start, the job of
>>>> dealing
>>>> with
>>>> the cluster was given to a co-worker. He has since moved on. I
>>>> know
>>>> just
>>>> enough about VMWare ESXI to keep the thing working. My new boss
>>>> wants to
>>>> get rid of VMWare and re-install everything on the bare metal
>>>> machines.
>>>>
>>>>
>>>> The VMWare host has 4 ethernet cables running to the switch. But
>>>> there
>>>> is only 1 virtual network port on the Linux virtual machine.
>>>> However,
>>>> lspci shows 32 "lines with VMware PCI Express Root" (whatever
>>>> that
>>>> is):
>>>>
>>>>
>>>> # lspci
>>>> 00:07.7 System peripheral: VMware Virtual Machine Communication
>>>> Interface (rev 10)
>>>> 00:10.0 SCSI storage controller: LSI Logic / Symbios Logic
>>>> 53c1030
>>>> PCI-X
>>>> Fusion-MPT Dual Ultra320 SCSI (rev 01)
>>>> 00:11.0 PCI bridge: VMware PCI bridge (rev 02)
>>>> 00:15.0 PCI bridge: VMware PCI Express Root Port (rev 01)
>>>> [...]
>>>> 00:18.7 PCI bridge: VMware PCI Express Root Port (rev 01)
>>>> 02:00.0 Ethernet controller: Intel Corporation 82545EM Gigabit
>>>> Ethernet
>>>> Controller (Copper) (rev 01)
>>>>
>>>>
>>>> The open-iscsi package is not installed on the Linux virtual
>>>> machine.
>>>> However, the ISCSI array shows up as /dev/sdb:
>>>>
>>>> # lsscsi
>>>> [2:0:0:0] disk VMware Virtual disk 1.0 /dev/sda
>>>> [2:0:1:0] disk EQLOGIC 100E-00 8.1 /dev/sdb
>>>>
>>>>
>>>> I'd kinda like to get the ISCSI array connected to a new bare
>>>> metal
>>>> Linux server w/o losing everybody's files. Do you think I can
>>>> just
>>>> follow the various hotos out there on connecting an ISCSI array
>>>> w/o
>>>> too
>>>> much trouble?
>>>> ------------------------------------------------------------------------
>>>> Ale mailing list
>>>> Ale at ale.org
>>>> https://mail.ale.org/mailman/listinfo/ale
>>>> <https://mail.ale.org/mailman/listinfo/ale>
>>>> See JOBS, ANNOUNCE and SCHOOLS lists at
>>>> http://mail.ale.org/mailman/listinfo
>>>> <http://mail.ale.org/mailman/listinfo>
>>>>
>>>>
>>>> --
>>>> Computers amplify human error
>>>> Super computers are really cool
>>> _______________________________________________
>>> Ale mailing list
>>> Ale at ale.org
>>> https://mail.ale.org/mailman/listinfo/ale
>>> See JOBS, ANNOUNCE and SCHOOLS lists at
>>> http://mail.ale.org/mailman/listinfo
>>>
>>
>
--
Derek Atkins 617-623-3745
derek at ihtfp.com www.ihtfp.com
Computer and Internet Security Consultant
More information about the Ale
mailing list