From shmookey at shmookey.net  Mon Feb  1 14:16:39 2010
From: shmookey at shmookey.net (Luke Williams)
Date: Mon, 1 Feb 2010 14:16:39 +0800
Subject: [tech] membership enquiry
In-Reply-To: <4B666744.6020508@maths.uwa.edu.au>
References: <4B666744.6020508@maths.uwa.edu.au>
Message-ID: <bfa572cf1001312216m7851530fl71a8d2c975db8a30@mail.gmail.com>

Hi Simon,

If your VPN problems include being kicked off constantly then the
chances are very good that the problem is with the software
configuration on your computer or a hardware error (at the wireless
layer), and it is unlikely that you will find our VPN to be any more
stable. Additionally, the internet connection you'll get through our
VPN is pretty slow and with the new quota restrictions for students
(1gb/day) there is more reason than ever to try to fix your problem
rather than step around it. You might be able to find help at UCC with
your issue, but you'll probably have better luck visiting the Student
Internet Support Office at either the Reid or Science library, who are
much better equipped (and paid) to fix these kind of issues.

That said, I certainly don't want to talk you out of joining our club!
If you are interested in joining I encourage you to come along to our
stall on O'Day - we have a $5 first-time member special for guild
members.

Regards,

Luke Williams
UCC Wheel Member

On Mon, Feb 1, 2010 at 1:31 PM, Simon Guest <simon at maths.uwa.edu.au> wrote:
> Hello,
>
> Apologies for the group email. I have seen your website and I think I would
> like to join if possible. I have been trying to connect my laptop wirelessly
> to the VPN SNAP network, and have been getting kicked off constantly. I
> understand you guys have a separate VPN server, and also the capability to
> ssh to machines your, without the need to go through VPN snap. Is this
> possible to do for lightish web browsing etc? I tried to drop by the UCC but
> the door was locked. If someone could let me know if I would be able to
> join, and also if someone could open the door for me this afternoon sometime
> after 4, then I would be most graeful.
>
> Many thanks
> Simon
>
>
>

From mattman at ucc.gu.uwa.edu.au  Tue Feb  2 14:36:22 2010
From: mattman at ucc.gu.uwa.edu.au (Matt Didcoe)
Date: Tue, 02 Feb 2010 14:36:22 +0800
Subject: [tech] RESNet speeds
Message-ID: <453c080250ba73dfa68944f9c10235bf@ucc.gu.uwa.edu.au>

Hey all,

Further to Richard Wotherspoon's email last Friday about the packet shaper
being removed, progress is being made to obtain the new equipment and ITS
hope to have it all in place by the time semester starts on February 22nd -
this is still up in the air though as vendors are experiencing delays
getting equipment to UWA.

At this stage, as you're all aware, the link has therefore been dropped
back to the original slow speeds we were experiencing before the trial.

This doesn't appear to have caused any major issues other than a bit of a
slow down when browsing the net at the club (but it's been quiet anyway...)

Further updates to come as we hear more from ITS.

Cheers
-- 
Matt Didcoe [MRD]
UCC Wheel Member
mattman at ucc.gu.uwa.edu.au

From shmookey at shmookey.net  Wed Feb  3 13:47:01 2010
From: shmookey at shmookey.net (Luke Williams)
Date: Wed, 3 Feb 2010 13:47:01 +0800
Subject: [tech] Fwd: Email changes for ironport change over
In-Reply-To: <201002031130.48758.toivo.pedaste@uwa.edu.au>
References: <201002031130.48758.toivo.pedaste@uwa.edu.au>
Message-ID: <bfa572cf1002022147v45fb6ff1id019f125853c0d94@mail.gmail.com>

---------- Forwarded message ----------
From: Toivo Pedaste <toivo.pedaste at uwa.edu.au>
Date: Wed, Feb 3, 2010 at 11:30 AM
Subject: Email changes for ironport change over
To: shmookey at ucc.asn.au


As part of the change over to the ?ironport gateways we are getting rid of the
policy routing and requiring all mail servers to explicitly send their
outgoing mail to antivirus.uwa.edu.au

The UCC host are currently the biggest ones that haven't changed over.

2923 ? ?dagon.rcpt.to[130.95.13.14]
1126 ? ?mooneye.ucc.gu.uwa.edu.au[130.95.13.9]

It would be good if the same could be done for the guild

786 ? ? guru.guild.uwa.edu.au[130.95.100.6]

--
Toivo Pedaste ? ? ? ? ? ? ?M463 ? ? ?Email: ?toivo.pedaste at uwa.edu.au
Information Technology Services, ? ? Phone: ?+61 8 6488 2605
University of Western Australia ? ? ?Fax: ? ?+61 8 6488 4751
"The time has come", the Walrus said, "to talk of many things"...

From matt at ucc.asn.au  Thu Feb  4 19:49:36 2010
From: matt at ucc.asn.au (Matt Johnston)
Date: Thu, 4 Feb 2010 19:49:36 +0800
Subject: [tech] Fwd: Email changes for ironport change over
In-Reply-To: <bfa572cf1002022147v45fb6ff1id019f125853c0d94@mail.gmail.com>
References: <201002031130.48758.toivo.pedaste@uwa.edu.au>
	<bfa572cf1002022147v45fb6ff1id019f125853c0d94@mail.gmail.com>
Message-ID: <20100204114936.GU16747@ucc.gu.uwa.edu.au>

I've sorted this out by adding

relayhost = antivirus.uwa.edu.au

to /etc/postfix/main.cf. It seems to work. I don't know who
runs the guild stuff, either Dave Cake or James Devenish?

Cheers,
Matt

On Wed, Feb 03, 2010 at 01:47:01PM +0800, Luke Williams wrote:
> ---------- Forwarded message ----------
> From: Toivo Pedaste <toivo.pedaste at uwa.edu.au>
> Date: Wed, Feb 3, 2010 at 11:30 AM
> Subject: Email changes for ironport change over
> To: shmookey at ucc.asn.au
> 
> 
> As part of the change over to the ?ironport gateways we are getting rid of the
> policy routing and requiring all mail servers to explicitly send their
> outgoing mail to antivirus.uwa.edu.au
> 
> The UCC host are currently the biggest ones that haven't changed over.
> 
> 2923 ? ?dagon.rcpt.to[130.95.13.14]
> 1126 ? ?mooneye.ucc.gu.uwa.edu.au[130.95.13.9]
> 
> It would be good if the same could be done for the guild
> 
> 786 ? ? guru.guild.uwa.edu.au[130.95.100.6]
> 
> --
> Toivo Pedaste ? ? ? ? ? ? ?M463 ? ? ?Email: ?toivo.pedaste at uwa.edu.au
> Information Technology Services, ? ? Phone: ?+61 8 6488 2605
> University of Western Australia ? ? ?Fax: ? ?+61 8 6488 4751
> "The time has come", the Walrus said, "to talk of many things"...

From mitch at ucc.gu.uwa.edu.au  Sat Feb  6 00:30:45 2010
From: mitch at ucc.gu.uwa.edu.au (Mitch Kelly)
Date: Sat, 6 Feb 2010 00:30:45 +0800
Subject: [tech] Changes to camwhore - New Tape library
Message-ID: <003901caa680$926a1190$b73e34b0$@gu.uwa.edu.au>

Hi All,

 

Camwhore:

I'm looking to decommission camwhore as the camera server.. Well not
decommission but virtualize camwhore on mylah, Not only does it have the 10k
SCSI Drives it now has a LTO-2 Tape library connected to it with capability
for 20 Tapes. (More with Stack)

 

The idea is to use 100~gb of mylah's 1TB RAID10 storage for camwhore, And
archive off to 2 of the tapes (800GB), Seeing as they are Images they should
compress rather well.

 

I will be in UCC over the weekend to plan and possibly do this.

 

 

Tape Library:

As mentioned in the minutes I have acquired a LTO-2 Tape library with space
for 20 Tapes (7TB), The library can handle LTO-3/4 Drives I believe,
Bringing this closer to 30TB But cost prohibits this for the time being, and
there is no real need for that much space.

 

The library will be populated with tapes over the coming weeks, But if you
have some spare LTO-2 Tapes, by all means leave them in the MR and we will
use them..

 

At the moment the library is connected to Mylah and powered up, Mylah has
standard SSH keys for wheel members.

 

Mitch

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20100206/fd6a2946/attachment.htm 

From adrian at ucc.gu.uwa.edu.au  Sat Feb  6 08:09:22 2010
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Sat, 6 Feb 2010 08:09:22 +0800
Subject: [tech] Changes to camwhore - New Tape library
In-Reply-To: <003901caa680$926a1190$b73e34b0$@gu.uwa.edu.au>
References: <003901caa680$926a1190$b73e34b0$@gu.uwa.edu.au>
Message-ID: <20100206000921.GA15610@ucc.gu.uwa.edu.au>

On Sat, Feb 06, 2010, Mitch Kelly wrote:

> The idea is to use 100~gb of mylah's 1TB RAID10 storage for camwhore, And
> archive off to 2 of the tapes (800GB), Seeing as they are Images they should
> compress rather well.

The images are already compressed? They won't (loss-less) compress much past that.




Adrian


From bob at ucc.gu.uwa.edu.au  Sat Feb  6 20:00:38 2010
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Sat, 6 Feb 2010 20:00:38 +0800 (WST)
Subject: [tech] Clubroom Machines
Message-ID: <alpine.DEB.1.10.1002061956070.28523@martello.ucc.gu.uwa.edu.au>

Hi,

As [TPG] and I were upgrading some components on cephalopod and combtail I 
noticed that combtails motherboard is on the way out with bulging power 
regulator caps. Due to this I will be swapping the hard disk and graphics 
card into the good Pentium D HP. If the old motherboard can be rapaired 
before it breaks I'm sure it would be a good spare.

Both machines will be keeping the same names, but will look significantly 
better.

Bob Adamson
UCC Wheel Member

|"Bureaucracy is a challenge to the be conquered with a righteous        |
|attitude, an intolerance for stupidity, and a bulldozer when necessary" |
| ---Peter's Laws                                                        |


From trs80 at ucc.gu.uwa.edu.au  Tue Feb  9 12:08:03 2010
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Tue, 9 Feb 2010 12:08:03 +0800 (WST)
Subject: [tech] mylah and musundo repurposing, virtualisation and storage
Message-ID: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>

[BOB] mentioned that he'd like to install Windows 7 on the desktops, which 
need a recent version of Samba to join the domain. We're currently running 
3.3.6 which isn't recent enough, and Blastwave's 3.4.3 version is broken: 
https://www.blastwave.org/forum/viewtopic.php?p=2095#p2095 OpenCSW only 
has 3.0.23, and SXCE b130 (which I upgraded musundo to yesterday) 3.0.37.

He then suggested moving Samba to a non-Solaris machine. A good candidate 
for this would be mylah, which could also take over amanda backups with 
the LTO2 tape library.

I've been wanting to have some sort of network storage for hosting VMs, 
which would ideally be running 64bit OpenSolaris to take advantage of ZFS 
deduplication for VM images and COMSTAR. Unfortunately mylah doesn't have 
AMD64 support, but if Samba, backups, /services and other bulky components 
of musundo:/space were moved to mylah then musundo and its fibre channel 
disk array could be used as the network storage server.

The disks of mylah and musundo are quite similar, 8x146GB SCSI and 18x73GB 
connected by one U160 SCSI bus and two 1Gb FC loops and currently 
configured as RAID6 and RAID10 (well, ZFS mirrors) respectively. Here's 
some benchmarks of the current setups:

Version  1.96       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
mylah           25G   142  99 75308  66 32936  19   330  98 104472  30 317.8  15
Latency             66344us     224ms     109ms   35798us   53357us     318ms

Version  1.96       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
musundo-b130    16G    43  98 49190  38 44650  34   100  98 180856  53  1153  52
Latency               603ms    3469ms    3454ms     271ms     208ms     222ms

As [AHC] has ranted about before, filesystems need a certain amount of 
empty space, and ZFS is particularly sensitive to this - here's a 
benchmark from when musundo:/space had only 26GB free:

Version  1.96       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
musundo-b130     8G    40  97 11209  10  9965   8    97  96 71121  21 226.9  13
Latency               548ms    3661ms    2542ms     402ms     906ms     977ms

There's still some open questions:
 o If we make mylah a Xen server, what OS should be used for the dom0? 
   Debian, which may have problems with multiple CPU domUs, or CentOS 5.4 
   which will be harder to maintain? Or OpenSolaris, which has its own 
   well-integrated Xen support and fancy network management (Crossbow)?
 o What virtualisation software do we use for the Sun servers from Arts, 
   when they arrive? Do we want to virtualise Windows Server 2008, which
   will require using VMWare, or dedicate an entire machine to it? Or run 
   VMWare on one, and Xen on the others? Again, if Xen, what dom0 OS?
 o Do we set up a separate network segment for storage traffic, which
   would require another gigabit switch, or is our main network 
   underutilised enough that it'd be overkill?

To summarise: Move samba, /services and backups to mylah, turn musundo 
into a reliable network storage server for VMs. It's a good short term 
solution that's also a decent long term solution if nothing more gets 
done. If no-one has any objections, I think we should do this before 
O'Day, with the downtime occurring at the cleanup.

-- 
# TRS-80              trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \
# UCC Wheel Member     http://trs80.ucc.asn.au/ #|  what squirrels do best     |
[ "There's nobody getting rich writing          ]|  -- Collect and hide your   |
[  software that I know of" -- Bill Gates, 1980 ]\  nuts." -- Acid Reflux #231 /

From nick at ucc.gu.uwa.edu.au  Tue Feb  9 13:53:09 2010
From: nick at ucc.gu.uwa.edu.au (Nick Bannon)
Date: Tue, 9 Feb 2010 13:53:09 +0800
Subject: [tech] mylah and musundo repurposing, virtualisation and	storage
In-Reply-To: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>
References: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>
Message-ID: <20100209055308.GB14786@ucc.gu.uwa.edu.au>

On Tue, Feb 09, 2010 at 12:08:03PM +0800, James Andrewartha wrote:
[...]
> To summarise: Move samba, /services and backups to mylah, turn musundo 
> into a reliable network storage server for VMs. [before O-Day]

That does sound good, thanks for the benchmarks.

> There's still some open questions:
>  o If we make mylah a Xen server, what OS should be used for the dom0? 
>    Debian, which may have problems with multiple CPU domUs, or CentOS 5.4 
>    which will be harder to maintain? Or OpenSolaris, which has its own 
>    well-integrated Xen support and fancy network management (Crossbow)?
>  o What virtualisation software do we use for the Sun servers from Arts, 
>    when they arrive? Do we want to virtualise Windows Server 2008, which
>    will require using VMWare, or dedicate an entire machine to it? Or run 
>    VMWare on one, and Xen on the others? Again, if Xen, what dom0 OS?

We're in the happy position that we have more decent hardware than we
know what to do with. It's nice to use a few platforms to let UCC members
learn about them, some for "production" and there can be others out in
the clubroom to experiment with and ponder upgrades with. I'm looking for
to the next round of handmedowns with amd64 and VT support, to make kvm
(and xenner, not so good at SMP either, yet) decent options.

I'd like to give Windows Server 2008 most of a machine, with VMWare,
it'll let us share the console and snapshot it nicely. I'd like to
have a couple of similar Xen machines, with working live migration.
Easy to get slack and never upgrade those, though, hence we might need
test machines in the clubroom that don't run all the time.

Whoever reinstalls mylah can choose, but try to get them to promise to
involve others in the regular maintenance.

>  o Do we set up a separate network segment for storage traffic, which
>    would require another gigabit switch, or is our main network 
>    underutilised enough that it'd be overkill?

I think we need a gigabit switch in each rack for cabling sanity.

I don't think we need a second just for SAN, yet, but it really might
be worth:
  * coloured SAN cables or cable tags;
  * physically placing the storage ports at one end of the switch; and
  * possibly not just throwing the traffic into the general VLAN trunk,
    instead physically patching a few storage ports between each switch.
    That would gain most of the real, tactile, "here's where the storage
    traffic flows, it's a bit less likely to explode if someone mistypes
    a config"

Nick.

-- 
   Nick Bannon   | "I made this letter longer than usual because
nick-sig at rcpt.to | I lack the time to make it shorter." - Pascal

From mitch at ucc.gu.uwa.edu.au  Tue Feb  9 17:48:13 2010
From: mitch at ucc.gu.uwa.edu.au (Mitch Kelly)
Date: Tue, 9 Feb 2010 17:48:13 +0800
Subject: [tech] Ubiquiti product demonstration night at UCC
Message-ID: <00c201caa96d$00049470$000dbd50$@gu.uwa.edu.au>

Hi Everyone,

 

Rob from freenet-antennas.com.au will be doing a presentation on Ubiquiti
products on the 26th February in the Cameron hall loft from 6:00PM till
8:00PM. Everyone is welcome but I do ask you RSVP to the event so we are
able to confirm numbers and organize space for it.

 

Some products on display and demonstration are:

Nanostation

Routerstation/RS Pro

Nanostation LoCo

Bullett Radios

Antennas

And many more.

 

Pizza will be eaten after the presentation, Most likely Broadway pizza,
Please bring $10 for Pizza if you wish to eat.

 

Hope to see you all there.

 

RSVP to mitch at ucc.asn.au

 

Mitch Kelly

UCC Wheel Member

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20100209/2b8ee43a/attachment.htm 

From adrian at ucc.gu.uwa.edu.au  Wed Feb 10 09:38:31 2010
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Wed, 10 Feb 2010 09:38:31 +0800
Subject: [tech] [ucc] Ubiquiti product demonstration night at UCC
In-Reply-To: <00c201caa96d$00049470$000dbd50$@gu.uwa.edu.au>
References: <00c201caa96d$00049470$000dbd50$@gu.uwa.edu.au>
Message-ID: <20100210013828.GB15610@ucc.gu.uwa.edu.au>


Cool. I'll come along with that. I've had some recent experience writing
and building FreeBSD firmware for this kit too.



adrian

On Tue, Feb 09, 2010, Mitch Kelly wrote:
> 
> Rob from freenet-antennas.com.au will be doing a presentation on Ubiquiti
> products on the 26th February in the Cameron hall loft from 6:00PM till
> 8:00PM. Everyone is welcome but I do ask you RSVP to the event so we are
> able to confirm numbers and organize space for it.
> 
>  
> 
> Some products on display and demonstration are:
> 
> Nanostation
> 
> Routerstation/RS Pro
> 
> Nanostation LoCo
> 
> Bullett Radios
> 
> Antennas
> 
> And many more.
> 
>  
> 
> Pizza will be eaten after the presentation, Most likely Broadway pizza,
> Please bring $10 for Pizza if you wish to eat.
> 
>  
> 
> Hope to see you all there.
> 
>  
> 
> RSVP to mitch at ucc.asn.au
> 
>  
> 
> Mitch Kelly
> 
> UCC Wheel Member
> 

From frenchie at ucc.gu.uwa.edu.au  Wed Feb 10 11:49:22 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Wed, 10 Feb 2010 11:49:22 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
Message-ID: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>

Hi Guys,

As has been discussed at a couple of meetings (at least one wheel
meeting and one committee meeting) Arts no longer need our four
original v20z servers so we're sending them UCCwards.

What's actually coming:
4x V20z's (1 with an old iLom version as it fails updates, 1 with a
dying/dead fan)
4x 73G 7200 rpm SCSI disks
4x Rail kits (and if they get put into the racks without these, I will
use them on the person responsible).
2x 36G 7200 rpm SCSI disks
2x single port, Qlogic Fibre Channel HBA's (multimode tranceiver built
into the card)
1x double port Qlogic FC HBA (requires SFPs)
1x 3 port Firewire card

All of the boxes have dual Opteron 244's (1.8GHz) with two having 4G
of ram and two with 3G. Matt Didcoe is currenly running DBAN on them
before we bring them over (mostly as I'm a lazy bastard). To my
knowledge, we've already earmarked these boxes as VM hosts, I'd
strongly recommend that we consolidate all the memory into two of them
and have the other two put aside for spare part cannibalism (esp as
they're 5 years old and parts are difficult to get).

Cheers,
Frenchie

From frenchie at ucc.gu.uwa.edu.au  Wed Feb 10 11:58:35 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Wed, 10 Feb 2010 11:58:35 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
Message-ID: <560b011002091958n2792ec98r4dea3abd4bb4bcc9@mail.gmail.com>

Oh and a spare, still wrapped in plastic PSU.

On 10 February 2010 11:49, James French <frenchie at ucc.gu.uwa.edu.au> wrote:
> Hi Guys,
>
> As has been discussed at a couple of meetings (at least one wheel
> meeting and one committee meeting) Arts no longer need our four
> original v20z servers so we're sending them UCCwards.
>
> What's actually coming:
> 4x V20z's (1 with an old iLom version as it fails updates, 1 with a
> dying/dead fan)
> 4x 73G 7200 rpm SCSI disks
> 4x Rail kits (and if they get put into the racks without these, I will
> use them on the person responsible).
> 2x 36G 7200 rpm SCSI disks
> 2x single port, Qlogic Fibre Channel HBA's (multimode tranceiver built
> into the card)
> 1x double port Qlogic FC HBA (requires SFPs)
> 1x 3 port Firewire card
>
> All of the boxes have dual Opteron 244's (1.8GHz) with two having 4G
> of ram and two with 3G. Matt Didcoe is currenly running DBAN on them
> before we bring them over (mostly as I'm a lazy bastard). To my
> knowledge, we've already earmarked these boxes as VM hosts, I'd
> strongly recommend that we consolidate all the memory into two of them
> and have the other two put aside for spare part cannibalism (esp as
> they're 5 years old and parts are difficult to get).
>
> Cheers,
> Frenchie
>

From trs80 at ucc.gu.uwa.edu.au  Wed Feb 10 20:53:29 2010
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Wed, 10 Feb 2010 20:53:29 +0800 (WST)
Subject: [tech] mylah and musundo repurposing, virtualisation and storage
In-Reply-To: <20100209055308.GB14786@ucc.gu.uwa.edu.au>
References: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>
	<20100209055308.GB14786@ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.1.10.1002100940190.6098@martello.ucc.gu.uwa.edu.au>

On Tue, 9 Feb 2010, Nick Bannon wrote:

> On Tue, Feb 09, 2010 at 12:08:03PM +0800, James Andrewartha wrote:
> [...]
> > There's still some open questions:
> >  o If we make mylah a Xen server, what OS should be used for the dom0? 
> >    Debian, which may have problems with multiple CPU domUs, or CentOS 5.4 
> >    which will be harder to maintain? Or OpenSolaris, which has its own 
> >    well-integrated Xen support and fancy network management (Crossbow)?
> >  o What virtualisation software do we use for the Sun servers from Arts, 
> >    when they arrive? Do we want to virtualise Windows Server 2008, which
> >    will require using VMWare, or dedicate an entire machine to it? Or run 
> >    VMWare on one, and Xen on the others? Again, if Xen, what dom0 OS?
> 
> We're in the happy position that we have more decent hardware than we
> know what to do with. It's nice to use a few platforms to let UCC members
> learn about them, some for "production" and there can be others out in
> the clubroom to experiment with and ponder upgrades with. I'm looking for
> to the next round of handmedowns with amd64 and VT support, to make kvm
> (and xenner, not so good at SMP either, yet) decent options.

The first such handmedown may arrive soon - troppo is being retired and 
donated to UCC. It's a Q6600 with VT and 8GB of RAM.
http://clubtroppo.com.au/2007/08/08/update-on-the-new-server/#more-3968

> I'd like to give Windows Server 2008 most of a machine, with VMWare,
> it'll let us share the console and snapshot it nicely. 

Yeah, [MRD] seems keen to run it under VMWare too.

> I'd like to
> have a couple of similar Xen machines, with working live migration.
> Easy to get slack and never upgrade those, though, hence we might need
> test machines in the clubroom that don't run all the time.

Live migration does require machines with CPUs from the same manufacturer 
(and ideally the same model) which may be a challenge with hetrogenous 
handmedowns. Although, now that I think about it, mermaid and mylah have 
similar enough CPUs that it should be feasible, but their differing 
versions of Xen could be a problem.

> Whoever reinstalls mylah can choose, but try to get them to promise to
> involve others in the regular maintenance.

CentOS seems to be the way to go, as OpenSolaris just integrated Xen 3.4 
so it's probably a bit early for production use. The only trick will be 
PCI passthrough of the SCSI controller for amanda.

> I think we need a gigabit switch in each rack for cabling sanity.
>
> I don't think we need a second just for SAN, yet, but it really might
> be worth:
>   * coloured SAN cables or cable tags;
>   * physically placing the storage ports at one end of the switch; and
>   * possibly not just throwing the traffic into the general VLAN trunk,
>     instead physically patching a few storage ports between each switch.
>     That would gain most of the real, tactile, "here's where the storage
>     traffic flows, it's a bit less likely to explode if someone mistypes
>     a config"

+1

> That does sound good, thanks for the benchmarks.

Here's another, with a 36GB FC drive used as a separate intent log:

Version  1.96       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
musundo-b130    16G    40  98 71699  59 51738  41   101  99 190597  56  1666  62
Latency               569ms    5398ms    3647ms     152ms     190ms     188ms

which shows better performance despite the 36GB drive being slower than 
all the other drives.

musundo will also be getting a second CPU board with another 2 1.2GHz 
US-III+ SPARCs and 8GB of RAM.

-- 
# TRS-80              trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \
# UCC Wheel Member     http://trs80.ucc.asn.au/ #|  what squirrels do best     |
[ "There's nobody getting rich writing          ]|  -- Collect and hide your   |
[  software that I know of" -- Bill Gates, 1980 ]\  nuts." -- Acid Reflux #231 /

From adrian at ucc.gu.uwa.edu.au  Thu Feb 11 08:13:01 2010
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Thu, 11 Feb 2010 08:13:01 +0800
Subject: [tech] mylah and musundo repurposing, virtualisation and	storage
In-Reply-To: <alpine.DEB.1.10.1002100940190.6098@martello.ucc.gu.uwa.edu.au>
References: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>
	<20100209055308.GB14786@ucc.gu.uwa.edu.au>
	<alpine.DEB.1.10.1002100940190.6098@martello.ucc.gu.uwa.edu.au>
Message-ID: <20100211001300.GC15610@ucc.gu.uwa.edu.au>

On Wed, Feb 10, 2010, James Andrewartha wrote:

> CentOS seems to be the way to go, as OpenSolaris just integrated Xen 3.4 
> so it's probably a bit early for production use. The only trick will be 
> PCI passthrough of the SCSI controller for amanda.

I suggest perhaps ignoring trying to do anything tricky with Xen to start
with and instead just migrate the non-tricky VMs into suitable places.

Tricky stuff can come later. Eg - it's a tape device, there's no real
reason why you couldn't run or write/run some form of tape proxy for
Amanda via dom0.




Adrian

From mitch at ucc.gu.uwa.edu.au  Thu Feb 11 08:15:10 2010
From: mitch at ucc.gu.uwa.edu.au (Mitch Kelly)
Date: Thu, 11 Feb 2010 08:15:10 +0800
Subject: [tech] mylah and musundo repurposing, virtualisation and	storage
In-Reply-To: <20100211001300.GC15610@ucc.gu.uwa.edu.au>
References: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>	<20100209055308.GB14786@ucc.gu.uwa.edu.au>	<alpine.DEB.1.10.1002100940190.6098@martello.ucc.gu.uwa.edu.au>
	<20100211001300.GC15610@ucc.gu.uwa.edu.au>
Message-ID: <000301caaaaf$469d8ed0$d3d8ac70$@gu.uwa.edu.au>

Dodgy hackery is why we are where we are today :P When only 1 or 2 people
know what the heck is going on :P


-----Original Message-----
From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
On Behalf Of Adrian Chadd
Sent: Thursday, February 11, 2010 8:13 AM
To: James Andrewartha
Cc: tech at ucc.gu.uwa.edu.au
Subject: Re: [tech] mylah and musundo repurposing, virtualisation and
storage

On Wed, Feb 10, 2010, James Andrewartha wrote:

> CentOS seems to be the way to go, as OpenSolaris just integrated Xen 3.4 
> so it's probably a bit early for production use. The only trick will be 
> PCI passthrough of the SCSI controller for amanda.

I suggest perhaps ignoring trying to do anything tricky with Xen to start
with and instead just migrate the non-tricky VMs into suitable places.

Tricky stuff can come later. Eg - it's a tape device, there's no real
reason why you couldn't run or write/run some form of tape proxy for
Amanda via dom0.




Adrian


From trs80 at ucc.gu.uwa.edu.au  Mon Feb 15 00:45:34 2010
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Mon, 15 Feb 2010 00:45:34 +0800 (WST)
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
Message-ID: <alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>

On Wed, 10 Feb 2010, James French wrote:

> As has been discussed at a couple of meetings (at least one wheel
> meeting and one committee meeting) Arts no longer need our four
> original v20z servers so we're sending them UCCwards.

[MRD] and I had a shot at installing VMWare ESXi 4 and Windows Server 
2008 R2 on one today. The VMWare install went OK, but not much else did:

> What's actually coming:
> 4x Rail kits (and if they get put into the racks without these, I will
> use them on the person responsible).

The rails are too long for all of our racks.

> 2x 36G 7200 rpm SCSI disks
> 2x single port, Qlogic Fibre Channel HBA's (multimode tranceiver built
> into the card)
> 1x double port Qlogic FC HBA (requires SFPs)
> 1x 3 port Firewire card

1x CD-ROM drive

Which makes installing off DVDs kind of hard. It didn't seem we could boot 
the VM from the virtual DVD drive from another computer, so we ended up 
ripping the disc to an .iso and using that, which took half an hour to 
upload to the local storage on the VM.

> All of the boxes have dual Opteron 244's (1.8GHz) with two having 4G

The CPUs are unfortunately revision CG, which makes them just slightly too 
old to work for 64bit guest virtualisation: http://kb.vmware.com/kb/1003945

So we were unable to install Windows Server 2008 R2, as it's only 
available in 64bit. We could just install it without virtualisation, but 
the install media is a DVD, so we can't boot from it, and the v20z doesn't 
boot from USB. [JCF] suggested using http://www.plop.at/en/bootmanager.html
as a bootloader for a USB drive with the installer copied to it using
http://www.intowindows.com/how-to-install-windows-7vista-from-usb-drive-detailed-100-working-guide/

-- 
# TRS-80              trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \
# UCC Wheel Member     http://trs80.ucc.asn.au/ #|  what squirrels do best     |
[ "There's nobody getting rich writing          ]|  -- Collect and hide your   |
[  software that I know of" -- Bill Gates, 1980 ]\  nuts." -- Acid Reflux #231 /

From frenchie at ucc.gu.uwa.edu.au  Tue Feb 16 11:09:27 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Tue, 16 Feb 2010 11:09:27 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
Message-ID: <560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>

> So we were unable to install Windows Server 2008 R2, as it's only
> available in 64bit. We could just install it without virtualisation, but
> the install media is a DVD, so we can't boot from it, and the v20z doesn't
> boot from USB. [JCF] suggested using http://www.plop.at/en/bootmanager.html
> as a bootloader for a USB drive with the installer copied to it using
> http://www.intowindows.com/how-to-install-windows-7vista-from-usb-drive-detailed-100-working-guide/

I have both of these prepared, I might take a long lunch and setup
Windows on one of the two hosts not intended for VMWare (keeping the
one remaining host as spare parts). I'd strongly reccomend that we
don't rush into AD and that we actually spend a bit of time thinking
about how we're going to do it properly.

F.

From adrian at ucc.gu.uwa.edu.au  Tue Feb 16 11:17:34 2010
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Tue, 16 Feb 2010 11:17:34 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
Message-ID: <20100216031734.GE15610@ucc.gu.uwa.edu.au>


I strongly recommend that people talk to other people who have
both deployed and supported AD setups in the past that have had
plenty of things go wrong AND have had to solve them.

Undoing a busted ass AD setup is .. not trivial. Migrating to
a newer small-scale NFS+NIS is a cinch in comparison.

2c,


Adrian

On Tue, Feb 16, 2010, James French wrote:
> 
> I have both of these prepared, I might take a long lunch and setup
> Windows on one of the two hosts not intended for VMWare (keeping the
> one remaining host as spare parts). I'd strongly reccomend that we
> don't rush into AD and that we actually spend a bit of time thinking
> about how we're going to do it properly.
> 
> F.

From frenchie at ucc.gu.uwa.edu.au  Tue Feb 16 11:19:24 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Tue, 16 Feb 2010 11:19:24 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <20100216031734.GE15610@ucc.gu.uwa.edu.au>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
	<20100216031734.GE15610@ucc.gu.uwa.edu.au>
Message-ID: <560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>

That would be me ;)

On 16 February 2010 11:17, Adrian Chadd <adrian at ucc.gu.uwa.edu.au> wrote:
>
> I strongly recommend that people talk to other people who have
> both deployed and supported AD setups in the past that have had
> plenty of things go wrong AND have had to solve them.
>
> Undoing a busted ass AD setup is .. not trivial. Migrating to
> a newer small-scale NFS+NIS is a cinch in comparison.
>
> 2c,
>
>
> Adrian
>
> On Tue, Feb 16, 2010, James French wrote:
>>
>> I have both of these prepared, I might take a long lunch and setup
>> Windows on one of the two hosts not intended for VMWare (keeping the
>> one remaining host as spare parts). I'd strongly reccomend that we
>> don't rush into AD and that we actually spend a bit of time thinking
>> about how we're going to do it properly.
>>
>> F.
>
>

From adrian at ucc.gu.uwa.edu.au  Tue Feb 16 11:20:48 2010
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Tue, 16 Feb 2010 11:20:48 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
	<20100216031734.GE15610@ucc.gu.uwa.edu.au>
	<560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>
Message-ID: <20100216032048.GF15610@ucc.gu.uwa.edu.au>

On Tue, Feb 16, 2010, James French wrote:
> That would be me ;)


I'm reading this as "frenchie is the new trs80." in terms of AD. :)

From frenchie at ucc.gu.uwa.edu.au  Tue Feb 16 11:29:23 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Tue, 16 Feb 2010 11:29:23 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <20100216032048.GF15610@ucc.gu.uwa.edu.au>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
	<20100216031734.GE15610@ucc.gu.uwa.edu.au>
	<560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>
	<20100216032048.GF15610@ucc.gu.uwa.edu.au>
Message-ID: <560b011002151929q71e5d0d1sb63bcff1f6837a8c@mail.gmail.com>

Well, to be quite honst, I'd also prefer it if we stuck with Samba as
it fits better with the tinkering nature of the club. AD is something
that's easy to fuck up when you have too many chefs in the kitchen.
See my first email

> I have both of these prepared, I might take a long lunch and setup
> Windows on one of the two hosts not intended for VMWare (keeping the
> one remaining host as spare parts). I'd strongly reccomend that we
> don't rush into AD and that we actually spend a bit of time thinking
> about how we're going to do it properly.

The subtle undertone was there about my reservations about AD :). That
said, a windows server probably _is_ something the club wants, so I'm
planning on building that all the same.

F.

On 16 February 2010 11:20, Adrian Chadd <adrian at ucc.gu.uwa.edu.au> wrote:
> On Tue, Feb 16, 2010, James French wrote:
>> That would be me ;)
>
>
> I'm reading this as "frenchie is the new trs80." in terms of AD. :)
>
>

From frenchie at ucc.gu.uwa.edu.au  Tue Feb 16 22:03:52 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Tue, 16 Feb 2010 22:03:52 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002151929q71e5d0d1sb63bcff1f6837a8c@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
	<20100216031734.GE15610@ucc.gu.uwa.edu.au>
	<560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>
	<20100216032048.GF15610@ucc.gu.uwa.edu.au>
	<560b011002151929q71e5d0d1sb63bcff1f6837a8c@mail.gmail.com>
Message-ID: <560b011002160603q42dee796r60e0ee3420dc535f@mail.gmail.com>

After a few hours of fiddling this afternoon, the closest I actually
got was half an install before it crapped out on a corrupted file. I'm
thinking I'll just lug it across campus to the office and do a
net-install there. Think the problems in installing are probably be
argument enough against running AD at this stage.

F.

On 16 February 2010 11:29, James French <frenchie at ucc.gu.uwa.edu.au> wrote:
> Well, to be quite honst, I'd also prefer it if we stuck with Samba as
> it fits better with the tinkering nature of the club. AD is something
> that's easy to fuck up when you have too many chefs in the kitchen.
> See my first email
>
>> I have both of these prepared, I might take a long lunch and setup
>> Windows on one of the two hosts not intended for VMWare (keeping the
>> one remaining host as spare parts). I'd strongly reccomend that we
>> don't rush into AD and that we actually spend a bit of time thinking
>> about how we're going to do it properly.
>
> The subtle undertone was there about my reservations about AD :). That
> said, a windows server probably _is_ something the club wants, so I'm
> planning on building that all the same.
>
> F.
>
> On 16 February 2010 11:20, Adrian Chadd <adrian at ucc.gu.uwa.edu.au> wrote:
>> On Tue, Feb 16, 2010, James French wrote:
>>> That would be me ;)
>>
>>
>> I'm reading this as "frenchie is the new trs80." in terms of AD. :)
>>
>>
>

From frenchie at ucc.gu.uwa.edu.au  Wed Feb 17 14:48:19 2010
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Wed, 17 Feb 2010 14:48:19 +0800
Subject: [tech] 4x Sunfire v20z's from FAHSS
In-Reply-To: <560b011002160603q42dee796r60e0ee3420dc535f@mail.gmail.com>
References: <560b011002091949p2196ceedyd1b73e5957ee067b@mail.gmail.com>
	<alpine.DEB.1.10.1002150034580.27171@martello.ucc.gu.uwa.edu.au>
	<560b011002151909o3c61ae2blbc09efb02215dab0@mail.gmail.com>
	<20100216031734.GE15610@ucc.gu.uwa.edu.au>
	<560b011002151919u291f07c2qce8d6c29593b8f96@mail.gmail.com>
	<20100216032048.GF15610@ucc.gu.uwa.edu.au>
	<560b011002151929q71e5d0d1sb63bcff1f6837a8c@mail.gmail.com>
	<560b011002160603q42dee796r60e0ee3420dc535f@mail.gmail.com>
Message-ID: <560b011002162248n7edba3c2q4602a10f75c17e7@mail.gmail.com>

Status Update:

Lugged the one I was playing with back to Arts and netinstalled win2k8
r2 STD but ran out of time before I got it  configured. Will finish up
tonight/tomorrow and then bring it back.

F.

On Tuesday, February 16, 2010, James French <frenchie at ucc.gu.uwa.edu.au> wrote:
> After a few hours of fiddling this afternoon, the closest I actually
> got was half an install before it crapped out on a corrupted file. I'm
> thinking I'll just lug it across campus to the office and do a
> net-install there. Think the problems in installing are probably be
> argument enough against running AD at this stage.
>
> F.
>
> On 16 February 2010 11:29, James French <frenchie at ucc.gu.uwa.edu.au> wrote:
>> Well, to be quite honst, I'd also prefer it if we stuck with Samba as
>> it fits better with the tinkering nature of the club. AD is something
>> that's easy to fuck up when you have too many chefs in the kitchen.
>> See my first email
>>
>>> I have both of these prepared, I might take a long lunch and setup
>>> Windows on one of the two hosts not intended for VMWare (keeping the
>>> one remaining host as spare parts). I'd strongly reccomend that we
>>> don't rush into AD and that we actually spend a bit of time thinking
>>> about how we're going to do it properly.
>>
>> The subtle undertone was there about my reservations about AD :). That
>> said, a windows server probably _is_ something the club wants, so I'm
>> planning on building that all the same.
>>
>> F.
>>
>> On 16 February 2010 11:20, Adrian Chadd <adrian at ucc.gu.uwa.edu.au> wrote:
>>> On Tue, Feb 16, 2010, James French wrote:
>>>> That would be me ;)
>>>
>>>
>>> I'm reading this as "frenchie is the new trs80." in terms of AD. :)
>>>
>>>
>>
>

From richard.wotherspoon at uwa.edu.au  Wed Feb 17 16:03:18 2010
From: richard.wotherspoon at uwa.edu.au (Richard Wotherspoon)
Date: Wed, 17 Feb 2010 16:03:18 +0800
Subject: [tech] [Resnet] Update
References: <94579C86-7E88-43C9-979E-4C967D26DB5C@staffad.uwa.edu.au>
Message-ID: <FDFE7AE7-5A15-4BD5-814A-FC943AB0F728@staffad.uwa.edu.au>

Hi All,

The new equipment for RESNet has now arrived and implementation has begun.

The majority of the work will be conducted during Thursday 18th (tomorrow) which will see intermittent outages to the RESNet link.

It is still expected that work will be completed by End of Business on Friday the 19th.

Please contact me directly if you have any concerns/questions.

Regards,
Richard


Richard Wotherspoon | Network Administrator

Information Technology Services, Information Services Division
The University of Western Australia
M463, 35 Stirling Highway, Crawley, WA, 6009
Telephone: +61 8 6488 2574
Fax: +61 8 6488 4751
CRICOS Code: 00126G

P Please consider the environment before printing this e-mail.

_______________________________________________
Resnet mailing list
Resnet at maillists.uwa.edu.au
http://maillists.uwa.edu.au/mailman/listinfo/resnet

From mattman at ucc.gu.uwa.edu.au  Thu Feb 18 08:00:47 2010
From: mattman at ucc.gu.uwa.edu.au (Matt Didcoe)
Date: Thu, 18 Feb 2010 08:00:47 +0800
Subject: [tech] Fwd: Re:  [Resnet] Update
Message-ID: <f23e5216e93859ea5629a9352f197692@ucc.gu.uwa.edu.au>

See below for some tweaks that have been made to our ResNet connection.

Also looks like we'll have blistering Internet for O'Day (so [BOB], you
can stop raging now :P)

-------- Original Message --------
Subject: Re: [tech] [Resnet] Update
Date: Thu, 18 Feb 2010 07:36:11 +0800
From: Richard Wotherspoon <richard.wotherspoon at uwa.edu.au>
To: Matt Didcoe <mattman at ucc.gu.uwa.edu.au>

Hey Matt,

I should really all be done today, EOB Friday is just a precaution in case
something goes terribly wrong!

One another note, the UCC currently has a deal for 128K minimum with a
bust to 1M on the RESNet pipe.


I have changed this slightly;
I have broken your 2 sets of subnets into their own classes instead of
sharing one 128K partition, as follows:

1) 10.203.13.0 /24 runs through "UCC_Private" with 128K minimum and burst
to 1M (This one includes any static NAT's you have)
2) 203.24.97.248 /29 runs through partition "UCC_Public" with 128K minimum
and burst to 1M (This one is the raw public space you have on VLAN 13)

I have done this to keep the rulesets clean (long story with the shapers)
but its a win for you guys anyway.

Cheers,
Richard




Richard Wotherspoon | Network Administrator

Information Technology Services, Information Services Division
The University of Western Australia
M463, 35 Stirling Highway, Crawley, WA, 6009
Telephone: +61 8 6488 2574
Fax: +61 8 6488 4751
CRICOS Code: 00126G

P Please consider the environment before printing this e-mail.


On 17/02/2010, at 6:34 PM, Matt Didcoe wrote:

Hi Richard,

If the outages are going to persist into Friday, could you drop us a quick
email just so we can plan ahead - we've got O'Day on Friday when we'll have
a number of new members coming through the club and the last thing we want
is a broken internet connection - if we know early enough we can change the
900 lines of iptables (ummm, spaghetti) to funnel everything (instead of
just on-net stuff) over the AARNET link instead.

Thanks mate :-)

Matt
University Computer Club


On 17/02/2010, at 4:03 PM, Richard Wotherspoon wrote:

Hi All,

The new equipment for RESNet has now arrived and implementation has begun.

The majority of the work will be conducted during Thursday 18th (tomorrow)
which will see intermittent outages to the RESNet link.

It is still expected that work will be completed by End of Business on
Friday the 19th.

Please contact me directly if you have any concerns/questions.

Regards,
Richard


Richard Wotherspoon | Network Administrator

Information Technology Services, Information Services Division
The University of Western Australia
M463, 35 Stirling Highway, Crawley, WA, 6009
Telephone: +61 8 6488 2574
Fax: +61 8 6488 4751
CRICOS Code: 00126G

P Please consider the environment before printing this e-mail.

_______________________________________________
Resnet mailing list
Resnet at maillists.uwa.edu.au<mailto:Resnet at maillists.uwa.edu.au>
http://maillists.uwa.edu.au/mailman/listinfo/resnet

-- 
Matt Didcoe [MRD]
UCC Wheel Member

From richard.wotherspoon at uwa.edu.au  Thu Feb 18 11:46:22 2010
From: richard.wotherspoon at uwa.edu.au (Richard Wotherspoon)
Date: Thu, 18 Feb 2010 11:46:22 +0800
Subject: [tech] [Resnet] Update
In-Reply-To: <FDFE7AE7-5A15-4BD5-814A-FC943AB0F728@staffad.uwa.edu.au>
References: <94579C86-7E88-43C9-979E-4C967D26DB5C@staffad.uwa.edu.au>
	<FDFE7AE7-5A15-4BD5-814A-FC943AB0F728@staffad.uwa.edu.au>
Message-ID: <825662F9-1F49-49EC-8A29-B8630DFC5211@staffad.uwa.edu.au>

Hi All,

The Traffic Management device has now been installed and link has been upgraded to 60M.

I have now re instated the percentage based split for each of the colleges like what was experienced during the trial period. (e.g 20% of the link is dedicated to CurrieHall etc.)

I have also created a general ruleset for the majority of internet protocols (i.e high priority for Skype, MSN Messenger over BitTorrent, etc) but if there are any reports of problems/slowness for other specific types, please let me know and i will prioritise accordingly.

Any questions/concerns please feel free to contact me,

Regards,
Richard




Richard Wotherspoon | Network Administrator

Information Technology Services, Information Services Division
The University of Western Australia
M463, 35 Stirling Highway, Crawley, WA, 6009
Telephone: +61 8 6488 2574
Fax: +61 8 6488 4751
CRICOS Code: 00126G

P Please consider the environment before printing this e-mail.

_______________________________________________
Resnet mailing list
Resnet at maillists.uwa.edu.au
http://maillists.uwa.edu.au/mailman/listinfo/resnet

From maset at ucc.asn.au  Thu Feb 18 14:21:41 2010
From: maset at ucc.asn.au (Anil Sharma)
Date: Thu, 18 Feb 2010 14:21:41 +0800
Subject: [tech] Multimeter
Message-ID: <22e859cb1002172221r1e222b35yfa642cfe8815b6e7@mail.gmail.com>

Is anyone planning on wanting to use the multimeter tonight? If not, I'm
going to theft it for a night.

Cheers,
Anil.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20100218/afe0f7ae/attachment.htm 

From trs80 at ucc.gu.uwa.edu.au  Tue Feb 23 21:13:04 2010
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Tue, 23 Feb 2010 21:13:04 +0800 (WST)
Subject: [tech] mylah and musundo repurposing, virtualisation and storage
In-Reply-To: <20100211001300.GC15610@ucc.gu.uwa.edu.au>
References: <alpine.DEB.1.10.1002090450060.7049@martello.ucc.gu.uwa.edu.au>
	<20100209055308.GB14786@ucc.gu.uwa.edu.au>
	<alpine.DEB.1.10.1002100940190.6098@martello.ucc.gu.uwa.edu.au>
	<20100211001300.GC15610@ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.1.10.1002231436150.30390@martello.ucc.gu.uwa.edu.au>

I've installed CentOS onto mylah, which is now known as maraena. I chose 
the name because CentOS is a whitebox version of RHEL, and maraena is a 
whitefish. Incidentally, muskellunge is the name that was going to be used 
for the Sun v20z running VMWare, as they eat other fish like a VM server 
eats other servers.

I'm using libvirt on maraena, which made installing mylah as a guest 
running Debian lenny a breeze:

virt-install -n mylah -r 2048 --vcpus=2 --os-type=linux --os-variant=debianlenny -l http://mirrors.uwa.edu.au/debian/dists/lenny/main/installer-i386/ --disk vol=maraena/mylah -w bridge:br1

In preparation for hosting /services and /away mylah is now running a 
2.6.32 kernel for ext4 and Samba 3.4.3 from backports.org.

On Thu, 11 Feb 2010, Adrian Chadd wrote:
> On Wed, Feb 10, 2010, James Andrewartha wrote:
> > CentOS seems to be the way to go, as OpenSolaris just integrated Xen 3.4 
> > so it's probably a bit early for production use. The only trick will be 
> > PCI passthrough of the SCSI controller for amanda.
> 
> I suggest perhaps ignoring trying to do anything tricky with Xen to start
> with and instead just migrate the non-tricky VMs into suitable places.

Yeah, I tried 3 different kernels while trying to get PCI passthrough 
working with a Debian guest, and they all crashed or hung.

The problem now is that the Xen networking scripts break the network 
configuration when they're run as part of the boot process, but not when 
run after the boot has finished. It turns out this is due to Xen and 
libvirt's network scripts conflicting, so I've set '(network-script 
/bin/true)' in /etc/xen/xend-config.sxp to disable the Xen scripts.
http://wiki.libvirt.org/page/Networking

So instead of using them, I've manually set up bridges of VLANs 2 (the 
default vlan on eth1) and 3 for the guest VMs. I've attached the relevant 
files from /etc/sysconfig/network-scripts. Guests on the clubroom network 
should use br1.3

A big help in debugging this was using the ability to use the DRAC's 
remote console from a terminal, not just the web Java applet. To do so, 
telnet to the DRAC and run "connect video" to use the text login on tty1. 

> Tricky stuff can come later. Eg - it's a tape device, there's no real
> reason why you couldn't run or write/run some form of tape proxy for
> Amanda via dom0.

That seems like a lot of effort compared to just running amanda in the 
dom0, and mounting /space/amanda from mylah for scratch space. 


-- 
# TRS-80              trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \
# UCC Wheel Member     http://trs80.ucc.asn.au/ #|  what squirrels do best     |
[ "There's nobody getting rich writing          ]|  -- Collect and hide your   |
[  software that I know of" -- Bill Gates, 1980 ]\  nuts." -- Acid Reflux #231 /
-------------- next part --------------
# Broadcom Corporation NetXtreme BCM5700 Gigabit Ethernet
VLAN=yes
DEVICE=eth1.3
BRIDGE=br1.3
ONBOOT=yes
-------------- next part --------------
# Broadcom Corporation NetXtreme BCM5700 Gigabit Ethernet
DEVICE=eth1
HWADDR=00:0B:DB:AD:BF:B0
BRIDGE=br1
ONBOOT=yes
-------------- next part --------------
# Broadcom Corporation NetXtreme BCM5700 Gigabit Ethernet
# Bridge for VMs
DEVICE=br1.3
TYPE=Bridge
ONBOOT=yes
BOOTPROTO=none
-------------- next part --------------
# Broadcom Corporation NetXtreme BCM5700 Gigabit Ethernet
# Bridge for VMs
DEVICE=br1
TYPE=Bridge
BOOTPROTO=static
IPADDR=130.95.13.40
NETMASK=255.255.255.192
ONBOOT=yes

From mitch at ucc.gu.uwa.edu.au  Wed Feb 24 07:58:48 2010
From: mitch at ucc.gu.uwa.edu.au (Mitch Kelly)
Date: Wed, 24 Feb 2010 07:58:48 +0800
Subject: [tech] FW: Special for UCC Members
Message-ID: <008901cab4e4$2471c760$6d555620$@gu.uwa.edu.au>



-----Original Message-----
From: Rob Clark [mailto:clark at freenet-antennas.com.au] 
Sent: Wednesday, February 24, 2010 7:23 AM
To: 'Mitch Kelly'; 'Mitch Kelly'
Subject: Special for UCC Members

Mitch,

Could you please let the UCC members know that if they want to, I am
happy to bring some NSM5s to the meeting on Friday and sell them as
follows:

Buy 1 at the 10+ price

That is they are $159.00 each incl GST and a Tax Invoice.

http://store.freenet-antennas.com.au/product_info.php?products_id=376 

Members need to aware that if deployed outside up a pole, they also need
a grounded CAT5 cable:

http://store.freenet-antennas.com.au/product_info.php?products_id=387

If members want to buy these on Friday - I need to know

- Name (for Invoice)
- Email address
- Qty of NSM5
- Qty of Shielded CAT5 cables

Regards

Rob




From jacques at chester.id.au  Wed Feb 24 08:03:00 2010
From: jacques at chester.id.au (Jacques Chester)
Date: Wed, 24 Feb 2010 08:03:00 +0800
Subject: [tech] troppo is down
Message-ID: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>

If there's anyone in the machine room, a reboot would be helpful.

Cheers,

JC.


From coxymla at gmail.com  Wed Feb 24 08:54:19 2010
From: coxymla at gmail.com (James Cox)
Date: Wed, 24 Feb 2010 08:54:19 +0800
Subject: [tech] troppo is down
In-Reply-To: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
References: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
Message-ID: <d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>

Considered a remote power switch?

There's one on OCAU right now for $165.
http://forums.overclockers.com.au/showthread.php?t=852410

On Wed, Feb 24, 2010 at 8:03 AM, Jacques Chester <jacques at chester.id.au> wrote:
> If there's anyone in the machine room, a reboot would be helpful.
>
> Cheers,
>
> JC.
>
>

From coxymla at gmail.com  Wed Feb 24 08:54:19 2010
From: coxymla at gmail.com (James Cox)
Date: Wed, 24 Feb 2010 08:54:19 +0800
Subject: [tech] troppo is down
In-Reply-To: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
References: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
Message-ID: <d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>

Considered a remote power switch?

There's one on OCAU right now for $165.
http://forums.overclockers.com.au/showthread.php?t=852410

On Wed, Feb 24, 2010 at 8:03 AM, Jacques Chester <jacques at chester.id.au> wrote:
> If there's anyone in the machine room, a reboot would be helpful.
>
> Cheers,
>
> JC.
>
>

From jacques at chester.id.au  Wed Feb 24 09:31:12 2010
From: jacques at chester.id.au (Jacques Chester)
Date: Wed, 24 Feb 2010 09:31:12 +0800
Subject: [tech] troppo is down
In-Reply-To: <d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>
References: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
	<d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>
Message-ID: <A5555A2C-5C2A-4776-A7B7-FE19FE4C2CBA@chester.id.au>


On 24/02/2010, at 8:54 AM, James Cox wrote:

> Considered a remote power switch?

An excellent idea. However troppo will be getting donated
to the club some time in the next few weeks, at which point
I won't be fussed by crashes :D

Cheers,

JC.


From mattman at ucc.gu.uwa.edu.au  Wed Feb 24 11:25:13 2010
From: mattman at ucc.gu.uwa.edu.au (Matt Didcoe)
Date: Wed, 24 Feb 2010 11:25:13 +0800
Subject: [tech] troppo is down
In-Reply-To: <d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>
References: <95B604A2-C83B-4D6D-BC27-910B5D6C28CE@chester.id.au>
	<d1381ac61002231654l76300e3ey74888b2947a7e8e9@mail.gmail.com>
Message-ID: <5cf77c0115544e67c12279719a57f9d1@ucc.gu.uwa.edu.au>

Restarted troppo - it seems happy :-)

On Wed, 24 Feb 2010 08:54:19 +0800, James Cox <coxymla at gmail.com> wrote:
> Considered a remote power switch?
> 
> There's one on OCAU right now for $165.
> http://forums.overclockers.com.au/showthread.php?t=852410
> 
> On Wed, Feb 24, 2010 at 8:03 AM, Jacques Chester <jacques at chester.id.au>
> wrote:
>> If there's anyone in the machine room, a reboot would be helpful.
>>
>> Cheers,
>>
>> JC.
>>
>>

-- 
Matt Didcoe [MRD]
UCC Wheel Member