From bob at ucc.gu.uwa.edu.au  Sun May  1 22:37:58 2011
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Sun, 1 May 2011 22:37:58 +0800 (WST)
Subject: [tech] Assorted remote management changes
Message-ID: <alpine.DEB.2.00.1105012223070.24588@motsugo.ucc.gu.uwa.edu.au>

Hi All,

I spent a little time tinkering with the IPMI on motsugo, and managed to 
get the interface and remote console working. I also set up the alert 
emails to go to hostmaster@ and tested it. A lot more of the temperature 
and voltage sensors work now, but I can't be bothered fixing up the stats 
page atm. Yay, now we can actually do something other than reboot remotely 
when things go pear shaped. That's only relevant to Wheel really but I 
figured some other people might be interested.

The process to get the IPMI working is fairly simple in Debian; install 
openipmi, lm-sensors, ipmitool and if you want, freeipmi-tools. You have 
to manually load the ipmi_devintf kernel module, and then /dev/ipmi0 
should appear. Then all the ipmi tools/interfaces should work.

Hopefully the same process will apply for murasoi, the replacement for 
madako.

Bob Adamson
UCC President

|"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 nick at ucc.gu.uwa.edu.au  Tue May  3 01:21:54 2011
From: nick at ucc.gu.uwa.edu.au (Nick Bannon)
Date: Tue, 3 May 2011 01:21:54 +0800
Subject: [tech] [wheel]  Assorted remote management changes
In-Reply-To: <alpine.DEB.2.00.1105012238050.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105012223070.24588@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105012238050.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <20110502172153.GK12203@ucc.gu.uwa.edu.au>

On Sun, May 01, 2011 at 10:49:54PM +0800, Bob Adamson wrote:
[...]
> Mylah's DRAC password is changed with the racadm utility, accessed by 
> telnetting to the DRAC with root/<rootpassword> and running the following 
> command 
> racadm config -g cfgUserAdmin -o cfgUserAdminPassword -i 1 "newpasswordhere"
> 
> In other news, another disk is on the way out on Mylah, judging by the 
> errors in dmesg. Film at 11.

Are there spare drives to hand?

Nick.

May  1 02:07:05 mylah kernel: [4713239.708119] sd 7:0:0:0: [sdg] Unhandled sense code
May  1 02:07:05 mylah kernel: [4713239.708175] sd 7:0:0:0: [sdg] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
May  1 02:07:05 mylah kernel: [4713239.708254] sd 7:0:0:0: [sdg] Sense Key : Medium Error [current] 
May  1 02:07:05 mylah kernel: [4713239.708328] Info fld=0x66c6e62
May  1 02:07:05 mylah kernel: [4713239.708367] sd 7:0:0:0: [sdg] Add.  Sense: Unrecovered read error
May  1 02:07:05 mylah kernel: [4713239.708432] sd 7:0:0:0: [sdg] CDB: Read(10): 28 00 06 6c 6c 3f 00 04 00 00
May  1 02:07:05 mylah kernel: [4713239.708607] end_request: I/O error, dev sdg, sector 107769442
May  1 02:07:05 mylah kernel: [4713239.764343] raid5:md1: read error corrected (8 sectors at 107769376 on sdg)

lrwxrwxrwx 1 root wheel 0 May  3 01:10 sdg -> ../devices/pci0000:01/0000:01:08.0/0000:02:06.1/host7/target7:0:0/7:0:0:0/block/sdg

Hmm, the logs don't seem to go back far enough, it looks like something
other than a Maxtor ATLAS10K4_146SCA.

  *-scsi:1
       description: SCSI storage controller
       product: AIC-7899P U160/m
       vendor: Adaptec
       physical id: 6.1
       bus info: pci at 0000:02:06.1
       logical name: scsi7
       version: 01
       width: 64 bits
       clock: 66MHz
       capabilities: scsi pm bus_master cap_list rom
       configuration: driver=aic7xxx latency=32 maxlatency=25 mingnt=40
       resources: irq:21 ioport:d800(size=256) memory:fcdfe000-fcdfefff
memory:f4100000-f411ffff(prefetchable)
     *-disk:0
          description: SCSI Disk
          physical id: 0.0.0
          bus info: scsi at 7:0.0.0
          logical name: /dev/sdg
          size: 136GiB (146GB)
          capabilities: partitioned partitioned:dos
          configuration: signature=ca3a57c3

-- 
   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 bob at ucc.gu.uwa.edu.au  Tue May  3 16:41:53 2011
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Tue, 3 May 2011 16:41:53 +0800 (WST)
Subject: [tech] [wheel]  Assorted remote management changes
In-Reply-To: <20110502172153.GK12203@ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105012223070.24588@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105012238050.24588@motsugo.ucc.gu.uwa.edu.au>
	<20110502172153.GK12203@ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105031640360.24588@motsugo.ucc.gu.uwa.edu.au>


Nope. We only have 73G disks spare, 146's are like hens teeth around here.

Bob

On Tue, 3 May 2011, Nick Bannon wrote:

> On Sun, May 01, 2011 at 10:49:54PM +0800, Bob Adamson wrote:
> [...]
> > Mylah's DRAC password is changed with the racadm utility, accessed by 
> > telnetting to the DRAC with root/<rootpassword> and running the following 
> > command 
> > racadm config -g cfgUserAdmin -o cfgUserAdminPassword -i 1 "newpasswordhere"
> > 
> > In other news, another disk is on the way out on Mylah, judging by the 
> > errors in dmesg. Film at 11.
> 
> Are there spare drives to hand?
> 
> Nick.
> 
> May  1 02:07:05 mylah kernel: [4713239.708119] sd 7:0:0:0: [sdg] Unhandled sense code
> May  1 02:07:05 mylah kernel: [4713239.708175] sd 7:0:0:0: [sdg] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
> May  1 02:07:05 mylah kernel: [4713239.708254] sd 7:0:0:0: [sdg] Sense Key : Medium Error [current] 
> May  1 02:07:05 mylah kernel: [4713239.708328] Info fld=0x66c6e62
> May  1 02:07:05 mylah kernel: [4713239.708367] sd 7:0:0:0: [sdg] Add.  Sense: Unrecovered read error
> May  1 02:07:05 mylah kernel: [4713239.708432] sd 7:0:0:0: [sdg] CDB: Read(10): 28 00 06 6c 6c 3f 00 04 00 00
> May  1 02:07:05 mylah kernel: [4713239.708607] end_request: I/O error, dev sdg, sector 107769442
> May  1 02:07:05 mylah kernel: [4713239.764343] raid5:md1: read error corrected (8 sectors at 107769376 on sdg)
> 
> lrwxrwxrwx 1 root wheel 0 May  3 01:10 sdg -> ../devices/pci0000:01/0000:01:08.0/0000:02:06.1/host7/target7:0:0/7:0:0:0/block/sdg
> 
> Hmm, the logs don't seem to go back far enough, it looks like something
> other than a Maxtor ATLAS10K4_146SCA.
> 
>   *-scsi:1
>        description: SCSI storage controller
>        product: AIC-7899P U160/m
>        vendor: Adaptec
>        physical id: 6.1
>        bus info: pci at 0000:02:06.1
>        logical name: scsi7
>        version: 01
>        width: 64 bits
>        clock: 66MHz
>        capabilities: scsi pm bus_master cap_list rom
>        configuration: driver=aic7xxx latency=32 maxlatency=25 mingnt=40
>        resources: irq:21 ioport:d800(size=256) memory:fcdfe000-fcdfefff
> memory:f4100000-f411ffff(prefetchable)
>      *-disk:0
>           description: SCSI Disk
>           physical id: 0.0.0
>           bus info: scsi at 7:0.0.0
>           logical name: /dev/sdg
>           size: 136GiB (146GB)
>           capabilities: partitioned partitioned:dos
>           configuration: signature=ca3a57c3
> 
> -- 
>    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.asn.au  Wed May  4 23:44:14 2011
From: mitch at ucc.asn.au (Mitch Kelly (UCC))
Date: Wed, 4 May 2011 23:44:14 +0800
Subject: [tech] Looking for a SAS Cable
In-Reply-To: <alpine.DEB.2.00.1104271249490.5713@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1104271249490.5713@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <005501cc0a72$1eb3a160$5c1ae420$@asn.au>

Hey,

Im hoping someone here can help me out, Im looking for an Internal SAS
Cable, Something like:
http://cgi.ebay.com.au/Amphenol-MiniSAS-4x-SAS-Internal-Cable-/320691911960?
pt=LH_DefaultDomain_0&hash=item4aaaba3d18#ht_500wt_1156

It must have the Mini-SAS on one end and the SAS+Power combined on the other
end.

Anyone?


From bob at ucc.gu.uwa.edu.au  Sat May  7 10:08:12 2011
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Sat, 7 May 2011 10:08:12 +0800 (WST)
Subject: [tech] Murasoi IP address
Message-ID: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>

I've been trying to work out why we're not using 130.95.13.1 as the 
address of our routing box, since that is generally the default gateway 
address (and what linux auto-configures when it doesn't have dhcp). As 
nobody on irc could give me an answer, I'm going to switch murasoi's IP to 
.1 . Apologies to anyone who now has to go and change ssh known_hosts, but 
I'm sure you can figure this out, being wheel members and all...

Bob

From trs80 at ucc.gu.uwa.edu.au  Sat May  7 15:47:19 2011
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Sat, 7 May 2011 15:47:19 +0800 (WST)
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>

On Sat, 7 May 2011, Bob Adamson wrote:

> I've been trying to work out why we're not using 130.95.13.1 as the 
> address of our routing box, since that is generally the default gateway 
> address (and what linux auto-configures when it doesn't have dhcp). As 
> nobody on irc could give me an answer, I'm going to switch murasoi's IP to 
> .1 . Apologies to anyone who now has to go and change ssh known_hosts, but 
> I'm sure you can figure this out, being wheel members and all...

10 hours on IRC, from midnight to 10am, is not sufficient notice for this 
sort of change. Please fuck off.

-- 
# 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 bob at ucc.gu.uwa.edu.au  Sat May  7 16:27:39 2011
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Sat, 7 May 2011 16:27:39 +0800 (WST)
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105071601360.24588@motsugo.ucc.gu.uwa.edu.au>

On Sat, 7 May 2011, James Andrewartha wrote:
> 10 hours on IRC, from midnight to 10am, is not sufficient notice for this 
> sort of change. Please fuck off.

Odd, I would have thought Frenchie would have mentioned to me why it was 
so hard to do. Regardless, I think your language leaves a lot to be 
desired.

The problem that has been pointed out to me is the effort required in 
changing over the default gateway on all the servers. I'm happy to do this 
once murasoi has been set up, and packet forwarding can be used so that 
it's not necessary to do it all at once.

I think the move should still happen, anything that isn't on dhcp but 
derives its config from an ip address assumes that the default gateway is 
.1 and will error if this is not the case. This frequently catches people 
when setting up machines on the machine room network. 

As Luke pointed out on irc, there's no reason that we have to both cut 
over the routing address and change the routing box at the same time. This 
seems like a wise idea, it can wait.

Bob Adamson
UCC President

|"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 mitch at ucc.asn.au  Sat May  7 16:31:12 2011
From: mitch at ucc.asn.au (Mitch Kelly (UCC))
Date: Sat, 7 May 2011 16:31:12 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <010101cc0c91$2593dba0$70bb92e0$@asn.au>

This kind of change should NOT be made without a wheel meeting. Or at least
14 days notice with prior consultation from wheel, I do agree with James,
Asking at midnight and using "Nobody could give you an answer", Allot of us
have Full time jobs, We have better things to do than sit on IRC at
midnight...

Not acceptable..

-----Original Message-----
From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
On Behalf Of James Andrewartha
Sent: Saturday, 7 May 2011 3:47 PM
To: Bob Adamson
Cc: tech at ucc.gu.uwa.edu.au
Subject: Re: [tech] Murasoi IP address

On Sat, 7 May 2011, Bob Adamson wrote:

> I've been trying to work out why we're not using 130.95.13.1 as the 
> address of our routing box, since that is generally the default gateway 
> address (and what linux auto-configures when it doesn't have dhcp). As 
> nobody on irc could give me an answer, I'm going to switch murasoi's IP to

> .1 . Apologies to anyone who now has to go and change ssh known_hosts, but

> I'm sure you can figure this out, being wheel members and all...

10 hours on IRC, from midnight to 10am, is not sufficient notice for this 
sort of change. Please fuck off.

-- 
# 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 mitch at ucc.asn.au  Sat May  7 16:33:06 2011
From: mitch at ucc.asn.au (Mitch Kelly (UCC))
Date: Sat, 7 May 2011 16:33:06 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071601360.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>	<alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105071601360.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <010501cc0c91$63502610$29f07230$@asn.au>

Set a sub interface.

Once everyone is changed over THEN take off the IP :)

-----Original Message-----
From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
On Behalf Of Bob Adamson
Sent: Saturday, 7 May 2011 4:28 PM
To: James Andrewartha
Cc: tech at ucc.gu.uwa.edu.au
Subject: Re: [tech] Murasoi IP address

On Sat, 7 May 2011, James Andrewartha wrote:
> 10 hours on IRC, from midnight to 10am, is not sufficient notice for this 
> sort of change. Please fuck off.

Odd, I would have thought Frenchie would have mentioned to me why it was 
so hard to do. Regardless, I think your language leaves a lot to be 
desired.

The problem that has been pointed out to me is the effort required in 
changing over the default gateway on all the servers. I'm happy to do this 
once murasoi has been set up, and packet forwarding can be used so that 
it's not necessary to do it all at once.

I think the move should still happen, anything that isn't on dhcp but 
derives its config from an ip address assumes that the default gateway is 
.1 and will error if this is not the case. This frequently catches people 
when setting up machines on the machine room network. 

As Luke pointed out on irc, there's no reason that we have to both cut 
over the routing address and change the routing box at the same time. This 
seems like a wise idea, it can wait.

Bob Adamson
UCC President

|"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 frenchie at ucc.gu.uwa.edu.au  Sat May  7 16:45:34 2011
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Sat, 7 May 2011 16:45:34 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071601360.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105071546420.21944@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105071601360.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <BANLkTik1LWjdZtPe-WLOv=cww7N6esg_4Q@mail.gmail.com>

> Odd, I would have thought Frenchie would have mentioned to me why it was
> so hard to do. Regardless, I think your language leaves a lot to be
> desired.

Don't drag me into this as your supporting witness. Nineteen words of
speculation for why we're using .3 does not constitute my support for
this change.

I fully endorse Trs80's closing statement.

F

From adrian at ucc.gu.uwa.edu.au  Sat May  7 17:30:26 2011
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Sat, 7 May 2011 17:30:26 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <20110507093025.GA22850@ucc.gu.uwa.edu.au>

On Sat, May 07, 2011, Bob Adamson wrote:
> I've been trying to work out why we're not using 130.95.13.1 as the 
> address of our routing box, since that is generally the default gateway 
> address (and what linux auto-configures when it doesn't have dhcp). As 
> nobody on irc could give me an answer, I'm going to switch murasoi's IP to 
> .1 . Apologies to anyone who now has to go and change ssh known_hosts, but 
> I'm sure you can figure this out, being wheel members and all...

I'm all for this, as long as Bob agrees to fix all the broken pieces.
(And I mean all. :)

:-)

If I ahd to guess as to why it's not .1, its because I bet the guild/UCS
was once .1 on that (or 130.95.100.0/23, the previous UCC network) many,
many moons ago.



Adrian


From nick at ucc.gu.uwa.edu.au  Sun May  8 09:26:34 2011
From: nick at ucc.gu.uwa.edu.au (Nick Bannon)
Date: Sun, 8 May 2011 09:26:34 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <20110507093025.GA22850@ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<20110507093025.GA22850@ucc.gu.uwa.edu.au>
Message-ID: <20110508012634.GQ12203@ucc.gu.uwa.edu.au>

On Sat, May 07, 2011 at 05:30:26PM +0800, Adrian Chadd wrote:
[...]
> If I ahd to guess as to why it's not .1, its because I bet the guild/UCS
> was once .1 on that (or 130.95.100.0/23, the previous UCC network) many,
> many moons ago.

Yep, it used to be our incoming network interface address:
http://lists.ucc.gu.uwa.edu.au/pipermail/tech/2001-May/001504.html

Makes sense to use 130.95.13.1 now, given a little warning here.

I was wondering how fast a TP-LINK TL-WR1043ND can do gigabit
routing. murasoi will do a better job, of course, but if for $75 we
could have a low power emergency backup router, that's a bonus. As you
mention it could even be a hot backup with CARP.

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 harrymc at decisions-and-designs.com.au  Sun May  8 12:19:12 2011
From: harrymc at decisions-and-designs.com.au (Harry McNally)
Date: Sun, 08 May 2011 12:19:12 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <20110508012634.GQ12203@ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>	<20110507093025.GA22850@ucc.gu.uwa.edu.au>
	<20110508012634.GQ12203@ucc.gu.uwa.edu.au>
Message-ID: <4DC619C0.4080808@decisions-and-designs.com.au>

On 08/05/11 09:26, Nick Bannon wrote:

> Makes sense to use 130.95.13.1 now, given a little warning here.
> 
> I was wondering how fast a TP-LINK TL-WR1043ND can do gigabit
> routing. murasoi will do a better job, of course, but if for $75 we
> could have a low power emergency backup router, that's a bonus. As you
> mention it could even be a hot backup with CARP.

Thanks for the moderate voice Nick :)

Would low power devices like the TL-WR1043ND ever be fast and flexible enough
for the main UCC router ?

Harry

From adrian at ucc.gu.uwa.edu.au  Sun May  8 13:47:53 2011
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Sun, 8 May 2011 13:47:53 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <4DC619C0.4080808@decisions-and-designs.com.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<20110507093025.GA22850@ucc.gu.uwa.edu.au>
	<20110508012634.GQ12203@ucc.gu.uwa.edu.au>
	<4DC619C0.4080808@decisions-and-designs.com.au>
Message-ID: <20110508054753.GB22850@ucc.gu.uwa.edu.au>

On Sun, May 08, 2011, Harry McNally wrote:
> On 08/05/11 09:26, Nick Bannon wrote:
> 
> > Makes sense to use 130.95.13.1 now, given a little warning here.
> > 
> > I was wondering how fast a TP-LINK TL-WR1043ND can do gigabit
> > routing. murasoi will do a better job, of course, but if for $75 we
> > could have a low power emergency backup router, that's a bonus. As you
> > mention it could even be a hot backup with CARP.
> 
> Thanks for the moderate voice Nick :)
> 
> Would low power devices like the TL-WR1043ND ever be fast and flexible enough
> for the main UCC router ?

Sort of[1].

Of course, that's the best case scenario. The minute someone loads it up
with traffic, or loads in too complicated an iptables ruleset, or the
iptables state tracking table fills the limited RAM available, or there's
a burst of small packets rather than large packets, the unit may fall over.

They're quite nice boxes though.

IMHO, a bit of investigation into some low spec core 2/i5 desktop board + 
RAM + CPU in a cheap case would likely yield a gigabit router suitable
for UCC, including all of the random frippery UCC tends to do on said
router, for under $800.



Adrian


From mitch at ucc.asn.au  Sun May  8 13:58:56 2011
From: mitch at ucc.asn.au (Mitch Kelly (UCC))
Date: Sun, 8 May 2011 13:58:56 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <4DC619C0.4080808@decisions-and-designs.com.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>	<20110507093025.GA22850@ucc.gu.uwa.edu.au>	<20110508012634.GQ12203@ucc.gu.uwa.edu.au>
	<4DC619C0.4080808@decisions-and-designs.com.au>
Message-ID: <01a101cc0d45$05c1e740$1145b5c0$@asn.au>

These kind of devices wont do "gigabit" routing, They max out at 60~Mbit
Routing speed, They will however do Switching at gigabit speed, So long as
it does not go via the CPU.


-----Original Message-----
From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
On Behalf Of Harry McNally
Sent: Sunday, 8 May 2011 12:19 PM
To: tech at ucc.gu.uwa.edu.au
Subject: Re: [tech] Murasoi IP address

On 08/05/11 09:26, Nick Bannon wrote:

> Makes sense to use 130.95.13.1 now, given a little warning here.
> 
> I was wondering how fast a TP-LINK TL-WR1043ND can do gigabit
> routing. murasoi will do a better job, of course, but if for $75 we
> could have a low power emergency backup router, that's a bonus. As you
> mention it could even be a hot backup with CARP.

Thanks for the moderate voice Nick :)

Would low power devices like the TL-WR1043ND ever be fast and flexible
enough
for the main UCC router ?

Harry


From blinken at gmail.com  Sun May  8 18:36:58 2011
From: blinken at gmail.com (Patrick Coleman)
Date: Sun, 8 May 2011 18:36:58 +0800
Subject: [tech] Murasoi IP address
In-Reply-To: <01a101cc0d45$05c1e740$1145b5c0$@asn.au>
References: <alpine.DEB.2.00.1105071002220.24588@motsugo.ucc.gu.uwa.edu.au>
	<20110507093025.GA22850@ucc.gu.uwa.edu.au>
	<20110508012634.GQ12203@ucc.gu.uwa.edu.au>
	<4DC619C0.4080808@decisions-and-designs.com.au>
	<01a101cc0d45$05c1e740$1145b5c0$@asn.au>
Message-ID: <-2089208409310624604@unknownmsgid>

Very little will route at a gigabit, if you want any kind of traffic
analysis or firewalling. This includes most server hardware with a
crappy chipset. Generally the issue is not the number of rules or
whatever (because state tracking does a good job of cutting down on
those) but the number of interrupts being raised by the NIC and so the
size of the packets coming through.

You can however get layer three switches for about $3-4k that will do
gigabit static routing with little to no analysis, if that's your
thing. Or a server with a fast (by MHz, not cores) CPU, an Intel (or
some Broadcom) chipsets that does interrupt coalescing sanely, and
probably a BSD OS.

As for the IP change, just make the new IP an alias on the interface,
change all the servers and devices and DNS, confirm with tcpdump that
there's no traffic crossing the interface for the old IP (for a good
24 hours), then remove the old IP. Doesn't have to take a week, or be
difficult, or require shouting.

Cheers,

Patrick

On 08/05/2011, at 13:59, "Mitch Kelly (UCC)" <mitch at ucc.asn.au> wrote:

> These kind of devices wont do "gigabit" routing, They max out at 60~Mbit
> Routing speed, They will however do Switching at gigabit speed, So long as
> it does not go via the CPU.
>
>
> -----Original Message-----
> From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
> On Behalf Of Harry McNally
> Sent: Sunday, 8 May 2011 12:19 PM
> To: tech at ucc.gu.uwa.edu.au
> Subject: Re: [tech] Murasoi IP address
>
> On 08/05/11 09:26, Nick Bannon wrote:
>
>> Makes sense to use 130.95.13.1 now, given a little warning here.
>>
>> I was wondering how fast a TP-LINK TL-WR1043ND can do gigabit
>> routing. murasoi will do a better job, of course, but if for $75 we
>> could have a low power emergency backup router, that's a bonus. As you
>> mention it could even be a hot backup with CARP.
>
> Thanks for the moderate voice Nick :)
>
> Would low power devices like the TL-WR1043ND ever be fast and flexible
> enough
> for the main UCC router ?
>
> Harry
>

From trs80 at ucc.gu.uwa.edu.au  Wed May 11 00:09:14 2011
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Wed, 11 May 2011 00:09:14 +0800 (WST)
Subject: [tech] NTP
Message-ID: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>

UCC's servers have pretty poor time at the moment - mussel was 6 minutes 
out, musundo something like half an hour. This is mostly because most 
servers are set to sync time from mooneye, which hasn't been running an 
NTP server recently. I restarted it, but it's not syncing with any 
machines. Obviously this is an issue for NFS, would anyone care to look 
into it some more?


-- 
# 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 matt at ucc.asn.au  Wed May 11 00:30:06 2011
From: matt at ucc.asn.au (Matt Johnston)
Date: Wed, 11 May 2011 00:30:06 +0800
Subject: [tech] NTP
In-Reply-To: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <20110510163005.GZ7103@ucc.gu.uwa.edu.au>

I'd recommend throwing away old ntpd and using chrony. It's
much saner syncing.

Matt

On Wed, May 11, 2011 at 12:09:14AM +0800, James Andrewartha wrote:
> UCC's servers have pretty poor time at the moment - mussel was 6 minutes 
> out, musundo something like half an hour. This is mostly because most 
> servers are set to sync time from mooneye, which hasn't been running an 
> NTP server recently. I restarted it, but it's not syncing with any 
> machines. Obviously this is an issue for NFS, would anyone care to look 
> into it some more?
> 
> 
> -- 
> # 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 bob at ucc.gu.uwa.edu.au  Wed May 11 00:45:16 2011
From: bob at ucc.gu.uwa.edu.au (Bob Adamson)
Date: Wed, 11 May 2011 00:45:16 +0800 (WST)
Subject: [tech] NTP
In-Reply-To: <20110510163005.GZ7103@ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
	<20110510163005.GZ7103@ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105110034050.24588@motsugo.ucc.gu.uwa.edu.au>

I'm no ntp expert, but from what I can tell, mooneye was looking at its 
hardware clock for the time, instead of time.uwa.edu.au. It's hardware 
clock is about a minute out, so I'm not quite sure what it would have 
done. Anyway, I added 'prefer' to the end of the time.uwa.edu.au line and 
restarted the server. It appears to have come good, but it will take a few 
days for it to drift if it's going to. As for the other servers, some 
didn't have ntp installed, others had the wrong/old servers configured, 
and some were matching mooneye's time. Most are matching now I think.

Bob


On Wed, 11 May 2011, Matt Johnston wrote:

> I'd recommend throwing away old ntpd and using chrony. It's
> much saner syncing.
> 
> Matt
> 
> On Wed, May 11, 2011 at 12:09:14AM +0800, James Andrewartha wrote:
> > UCC's servers have pretty poor time at the moment - mussel was 6 minutes 
> > out, musundo something like half an hour. This is mostly because most 
> > servers are set to sync time from mooneye, which hasn't been running an 
> > NTP server recently. I restarted it, but it's not syncing with any 
> > machines. Obviously this is an issue for NFS, would anyone care to look 
> > into it some more?
> > 
> > 
> > -- 
> > # 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 michael at deegan.id.au  Wed May 11 11:27:24 2011
From: michael at deegan.id.au (Michael Deegan)
Date: Wed, 11 May 2011 11:27:24 +0800
Subject: [tech] NTP
In-Reply-To: <alpine.DEB.2.00.1105110034050.24588@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
	<20110510163005.GZ7103@ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105110034050.24588@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <20110511032724.GI2537@deegan.id.au>

Morning,

On Wed, May 11, 2011 at 12:45:16AM +0800, Bob Adamson wrote:
> I'm no ntp expert, but from what I can tell, mooneye was looking at its 
> hardware clock for the time, instead of time.uwa.edu.au. It's hardware 

What I think can happen is on startup, ntp skips any servers it can't
resolve the IP for. Obviously this is a problem for machines that boot
while the network's down.

-MD

-- 
-------------------------------------------------------------------------------
Michael Deegan                Hugaholic                http://www.deegan.id.au/
--------------------- Guvax bs vg nf ribyhgvba va npgvba. ---------------------

From harrymc at ucc.gu.uwa.edu.au  Wed May 11 15:16:31 2011
From: harrymc at ucc.gu.uwa.edu.au (Harry)
Date: Wed, 11 May 2011 15:16:31 +0800
Subject: [tech] AVR evaluation hardware
Message-ID: <4DCA37CF.5080306@ucc.gu.uwa.edu.au>

Hello

Just got some Braemac news pointing to these:

http://www.atmel.com/products/AVR/xplain.asp?category_id=163&family_id=607

Arduino has a place but these may be useful alternatives for some AVR projects
(UCC or personal).

Harry

From zarquin at ucc.gu.uwa.edu.au  Wed May 11 15:58:10 2011
From: zarquin at ucc.gu.uwa.edu.au (Alwyn Lloyd)
Date: Wed, 11 May 2011 15:58:10 +0800 (WST)
Subject: [tech] NTP
In-Reply-To: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105111557380.21514@mussel.ucc.gu.uwa.edu.au>


So no-one found the GPS time reciever?

Sadface.

[ZAR]


> UCC's servers have pretty poor time at the moment - mussel was 6 minutes
> out, musundo something like half an hour. This is mostly because most
> servers are set to sync time from mooneye, which hasn't been running an
> NTP server recently. I restarted it, but it's not syncing with any
> machines. Obviously this is an issue for NFS, would anyone care to look
> into it some more?
>



From harrymc at decisions-and-designs.com.au  Wed May 11 16:43:40 2011
From: harrymc at decisions-and-designs.com.au (Harry McNally)
Date: Wed, 11 May 2011 16:43:40 +0800
Subject: [tech] NTP
In-Reply-To: <alpine.DEB.2.00.1105111557380.21514@mussel.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105110004230.23420@motsugo.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105111557380.21514@mussel.ucc.gu.uwa.edu.au>
Message-ID: <4DCA4C3C.3050600@decisions-and-designs.com.au>

On 11/05/11 15:58, Alwyn Lloyd wrote:
> 
> So no-one found the GPS time reciever?
> 
> Sadface.

Hello [ZAR]

I've asked while at UCC and no so far. I was interested in the mitch GPS
modules as an alternative time sync but the datasheet on those says the 1PPS
is not suitable for time sync. Seems odd because the time sync boards of old
accepted that the PPS signal would jitter. Not sure why this one would be
different.

[HMC]

>> UCC's servers have pretty poor time at the moment - mussel was 6 minutes
>> out, musundo something like half an hour. This is mostly because most
>> servers are set to sync time from mooneye, which hasn't been running an
>> NTP server recently. I restarted it, but it's not syncing with any
>> machines. Obviously this is an issue for NFS, would anyone care to look
>> into it some more?
>>
> 
> 
> 


From zanchey at ucc.gu.uwa.edu.au  Wed May 11 23:51:46 2011
From: zanchey at ucc.gu.uwa.edu.au (David Adam)
Date: Wed, 11 May 2011 23:51:46 +0800 (WST)
Subject: [tech] [ucc] Minutes of Meeting 6th May 2011
In-Reply-To: <alpine.DEB.2.00.1105062031280.31398@mussel.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105062031280.31398@mussel.ucc.gu.uwa.edu.au>
Message-ID: <alpine.DEB.2.00.1105112309270.7643@motsugo.ucc.gu.uwa.edu.au>

On Fri, 6 May 2011, Conrad Pogson wrote:

> UCC Committee Minutes 06-05-2011
> Meeting opened at 17:08
> 
> Present: [BOB] [LOL] [TPG] [ASH] [DJA] [CJS] [SZM] [EDO] [AHC]
> Apologies: [MRD]
> 
> General Business
> ----------------
> - [DAA] wants to purchase TP-LINK TL-WR1043ND wireless N AP ($75)
>   - [AHC] claims this is 'actually pretty shit' and the 'range is crap'

[AHC] actually recommended this model in a thread on committee@, but all 
intelligent people change their minds from time to time.

>   - Will apparently work inside the room, but not much more.
>   - Difference between this and the old AP, is that OpenWRT wont run on our
>     current AP and it's faster. We will be able to do byte counting on the new
>     AP.
>   - More discussion on tech@ would be good, deferred.

OpenWRT runs on our current AP. What does not run is a wireless driver 
that supports RADIUS accounting i.e. the easiest way to keep track of how 
much traffic people use. So if keeping track of this is important to the 
committee, we need a wireless AP that supports this, which includes many 
(but not all) devices that run OpenWRT, specifically those that use the 
hostapd toolkit.

The TL-WR1043ND has the advantage of having a Gigabit uplink available, 
which means that under ideal conditions the full bandwidth of the 802.11n 
specification would be available.

Other options include the Cisco Aironet 1040/1140-grade kit 
(probably around ~$500+), Linksys WRT160NL (not available new), Ubiquiti 
PowerAP N (no gigabit) or Ubiquiti UniFi AP ($120-165, no gigabit, not 
totally convinced of software support for accounting).

Committee should therefore consider:
 - do we want a new AP that does 802.11n;
 - do we care about accounting for wireless users; and
 - how much money do we want to spend if yes to either of the above?

Dual-band devices (which will offer better performance for some users) are 
definitely more expensive. Finding devices which are both dual-band and 
have Gigabit Ethernet is left as an exercise to the reader.

[DAA]

From blinken at gmail.com  Thu May 12 00:20:20 2011
From: blinken at gmail.com (Patrick Coleman)
Date: Thu, 12 May 2011 00:20:20 +0800
Subject: [tech] [ucc] Minutes of Meeting 6th May 2011
In-Reply-To: <alpine.DEB.2.00.1105112309270.7643@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105062031280.31398@mussel.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105112309270.7643@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <8407891720997608216@unknownmsgid>

Cheers,

Patrick

On 11/05/2011, at 23:51, David Adam <zanchey at ucc.gu.uwa.edu.au> wrote:

> On Fri, 6 May 2011, Conrad Pogson wrote:
>
>> UCC Committee Minutes 06-05-2011
>> Meeting opened at 17:08
>>
>> Present: [BOB] [LOL] [TPG] [ASH] [DJA] [CJS] [SZM] [EDO] [AHC]
>> Apologies: [MRD]
>>
>> General Business
>> ----------------
>> - [DAA] wants to purchase TP-LINK TL-WR1043ND wireless N AP ($75)
>>  - [AHC] claims this is 'actually pretty shit' and the 'range is crap'
>
> [AHC] actually recommended this model in a thread on committee@, but all
> intelligent people change their minds from time to time.
>
>>  - Will apparently work inside the room, but not much more.
>>  - Difference between this and the old AP, is that OpenWRT wont run on our
>>    current AP and it's faster. We will be able to do byte counting on the new
>>    AP.
>>  - More discussion on tech@ would be good, deferred.
>
> OpenWRT runs on our current AP. What does not run is a wireless driver
> that supports RADIUS accounting i.e. the easiest way to keep track of how
> much traffic people use. So if keeping track of this is important to the
> committee, we need a wireless AP that supports this, which includes many
> (but not all) devices that run OpenWRT, specifically those that use the
> hostapd toolkit.
>
> The TL-WR1043ND has the advantage of having a Gigabit uplink available,
> which means that under ideal conditions the full bandwidth of the 802.11n
> specification would be available.
>
> Other options include the Cisco Aironet 1040/1140-grade kit
> (probably around ~$500+)

If enterprise kit is an option, consider the Aruba Instant range,
which is their new thick-ap (they were previously thin only). Unsure
on pricing, but there wouldn't be too much change from $1000.

You would however get dual-band gigabit 2x2 MIMO, and all the RADIUS
accounting/guest management whatever you can poke a stick at, built
in. They have some really cool tech (the controller-based models will
do full spectrum analysis so you can spot interference, plus track the
physical locations of users, and spoof deauth packets to neighboring
APs to prevent your users associating), though I'm not sure what you
get in this model. Possibly just RADIUS with guest captive portal.

If ctte is seriously interested, I can ask our sales rep if a discount
or donation might be an option for UCC. Let me know.

-Patrick

From adrian at ucc.gu.uwa.edu.au  Thu May 12 15:43:37 2011
From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd)
Date: Thu, 12 May 2011 15:43:37 +0800
Subject: [tech] [ucc] Minutes of Meeting 6th May 2011
In-Reply-To: <alpine.DEB.2.00.1105112309270.7643@motsugo.ucc.gu.uwa.edu.au>
References: <alpine.DEB.2.00.1105062031280.31398@mussel.ucc.gu.uwa.edu.au>
	<alpine.DEB.2.00.1105112309270.7643@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <20110512074337.GA20058@ucc.gu.uwa.edu.au>

On Wed, May 11, 2011, David Adam wrote:
> > General Business
> > ----------------
> > - [DAA] wants to purchase TP-LINK TL-WR1043ND wireless N AP ($75)
> >   - [AHC] claims this is 'actually pretty shit' and the 'range is crap'
> 
> [AHC] actually recommended this model in a thread on committee@, but all 
> intelligent people change their minds from time to time.

I gave my local unit a bit more of a thrashing. The throughput dropped
off reasonably quickly outside of my small living room.

There's a dlink model which has been suggested to me - it's a dual 2/5ghz
band unit and it apparently has much better range and throughput.

> Dual-band devices (which will offer better performance for some users) are 
> definitely more expensive. Finding devices which are both dual-band and 
> have Gigabit Ethernet is left as an exercise to the reader.

Cue the D-Link DIR-825.

http://www.shopbot.com.au/pp-d-link-dir-825-price-141873.html

http://www.dlink.com/products/?pid=681

11n, 2/5ghz simultaneously (2+ath9k NICs), runs openwrt, 4+1 gige port,
USB2.

Openwrt wiki page: http://wiki.openwrt.org/toh/d-link/dir-825

On the flip-side it's around $200. But as I said, I've been told by
other wifi developers that this unit Definitely Doesn't Suck.


Adrian


From danielax at gmail.com  Thu May 12 20:08:39 2011
From: danielax at gmail.com (Daniel Axtens)
Date: Thu, 12 May 2011 20:08:39 +0800
Subject: [tech] [committee] 24 hour warning: This Week's Committee
	Meeting Agenda
In-Reply-To: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
Message-ID: <85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>

> we must discuss backups. Namely the offsite backups. Namely the offsite backups that don't exist. I motion that we compel wheel to setup offsite backups, possibly in arts.

So, tech@ would be the appropriate place to have this discussion.

[DJA]


From frenchie at ucc.gu.uwa.edu.au  Thu May 12 21:12:30 2011
From: frenchie at ucc.gu.uwa.edu.au (James French)
Date: Thu, 12 May 2011 21:12:30 +0800
Subject: [tech] [committee] 24 hour warning: This Week's Committee
	Meeting Agenda
In-Reply-To: <85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
	<85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>
Message-ID: <BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>

My suggestion would be to exchange manduba for musundo+disk at Arts
and do it ASAP. The understanding I put in place four* years ago was
agreed on by the late Mike Neville and was kept running under the last
FC manager (who left a month ago). There's going to be a new manager
in the not to distant future and there is no obligation for them keep
us in the room.

We'll probably want to avoid drawing attention to our stuff until the
new manager is settled and MRD has an opportunity to determine whether
or not they'll be amicable to helping us out. Basically what I'm
saying is if we want to make changes, do it now before the new guy
hits the ground and has opportunity to notice that things are
different all of a sudden.

MRD may wish to comment on this, as he's now me over there and I'm
really only saying anything as I coerced my old boss into letting us
do it in the first place.

F.

* !! is it really that long ago, I'm getting old?

On Thu, May 12, 2011 at 8:08 PM, Daniel Axtens <danielax at gmail.com> wrote:
>> we must discuss backups. Namely the offsite backups. Namely the offsite backups that don't exist. I motion that we compel wheel to setup offsite backups, possibly in arts.
>
> So, tech@ would be the appropriate place to have this discussion.
>
> [DJA]
>
>

From jacques at chester.id.au  Fri May 13 06:12:52 2011
From: jacques at chester.id.au (Jacques Chester)
Date: Fri, 13 May 2011 07:42:52 +0930
Subject: [tech] [committee] 24 hour warning: This Week's Committee
	Meeting Agenda
In-Reply-To: <BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
	<85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>
	<BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>
Message-ID: <7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>


On 12/05/2011, at 10:42 PM, James French wrote:

> My suggestion would be to exchange manduba for musundo+disk at Arts
> and do it ASAP. The understanding I put in place four* years ago was
> agreed on by the late Mike Neville and was kept running under the last
> FC manager (who left a month ago). There's going to be a new manager
> in the not to distant future and there is no obligation for them keep
> us in the room.

If Arts kick us out, I've found tarsnap to be cheap and effective
for backing up my Ozblogistan servers. I've got daily snapshots
going back about 9 months of ~4Gb of stuff. Total cost, $8.

Cheers,

JC.


From matt at didcoe.id.au  Fri May 13 08:21:53 2011
From: matt at didcoe.id.au (Matt Didcoe)
Date: Fri, 13 May 2011 08:21:53 +0800
Subject: [tech] [committee] 24 hour warning: This Week's Committee
	Meeting Agenda
In-Reply-To: <7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
	<85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>
	<BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>
	<7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>
Message-ID: <BANLkTi=DbfLy5dYzF4JNoFPVGdv5EDK_RQ@mail.gmail.com>

Offsite backups are great, but lets be a little more specific - WHAT
are UCC looking to backup offsite? That'll effect the amount of disk
required, how much it'll flog the network etc.

In theory Arts (read: Rob and I) have no problem with the idea and we
agree - sooner rather than later is a good idea, but there are a few
things to iron out - like what UCC would be dropping in and what
effect, if any, the core network upgrade will have (given the Arts DC
is served off a different VRF zone under the new plan).

I'll have a chat to Richard at IS about the second bit and perhaps we
could clarify here:

1) What we want to store offsite
2) What disk could be coming with musundo

Cheers,
Matt [MRD]

On Fri, May 13, 2011 at 6:12 AM, Jacques Chester <jacques at chester.id.au> wrote:
>
> On 12/05/2011, at 10:42 PM, James French wrote:
>
>> My suggestion would be to exchange manduba for musundo+disk at Arts
>> and do it ASAP. The understanding I put in place four* years ago was
>> agreed on by the late Mike Neville and was kept running under the last
>> FC manager (who left a month ago). There's going to be a new manager
>> in the not to distant future and there is no obligation for them keep
>> us in the room.
>
> If Arts kick us out, I've found tarsnap to be cheap and effective
> for backing up my Ozblogistan servers. I've got daily snapshots
> going back about 9 months of ~4Gb of stuff. Total cost, $8.
>
> Cheers,
>
> JC.
>
>

From mitch at ucc.asn.au  Fri May 13 10:35:13 2011
From: mitch at ucc.asn.au (Mitch Kelly (UCC))
Date: Fri, 13 May 2011 10:35:13 +0800
Subject: [tech] [committee] 24 hour warning: This Week's
	Committee	Meeting Agenda
In-Reply-To: <7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>	<85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>	<BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>
	<7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>
Message-ID: <000701cc1116$6964fb00$3c2ef100$@asn.au>

Im happy to give 100~GB space one of my colo's

-----Original Message-----
From: tech-bounces at ucc.gu.uwa.edu.au [mailto:tech-bounces at ucc.gu.uwa.edu.au]
On Behalf Of Jacques Chester
Sent: Friday, 13 May 2011 6:13 AM
To: tech
Subject: Re: [tech] [committee] 24 hour warning: This Week's Committee
Meeting Agenda


On 12/05/2011, at 10:42 PM, James French wrote:

> My suggestion would be to exchange manduba for musundo+disk at Arts
> and do it ASAP. The understanding I put in place four* years ago was
> agreed on by the late Mike Neville and was kept running under the last
> FC manager (who left a month ago). There's going to be a new manager
> in the not to distant future and there is no obligation for them keep
> us in the room.

If Arts kick us out, I've found tarsnap to be cheap and effective
for backing up my Ozblogistan servers. I've got daily snapshots
going back about 9 months of ~4Gb of stuff. Total cost, $8.

Cheers,

JC.


From oxinabox at ucc.asn.au  Fri May 13 13:40:08 2011
From: oxinabox at ucc.asn.au (oxinabox at ucc.asn.au)
Date: Fri, 13 May 2011 13:40:08 +0800
Subject: [tech] [committee] 24 hour warning: This Week's Committee
 Meeting Agenda
In-Reply-To: <BANLkTi=DbfLy5dYzF4JNoFPVGdv5EDK_RQ@mail.gmail.com>
References: <20110512090002.355136006B@motsugo.ucc.gu.uwa.edu.au>
	<85FFB853-3513-4470-BAD6-F07288978A5F@gmail.com>
	<BANLkTiko6qyN5yK-r7Qxw8up8Du9yBQ7gw@mail.gmail.com>
	<7EA79971-D9FF-4A90-BCB4-0ABBF08886DB@chester.id.au>
	<BANLkTi=DbfLy5dYzF4JNoFPVGdv5EDK_RQ@mail.gmail.com>
Message-ID: <209b17a2c786750704fbd00dc6efc44a@secure.ucc.asn.au>

> 1) What we want to store offsite

 I wasn't at the meeting, but in my mind the most important things to 
 have stored in any backup would be the source control repositories. (cos 
 wel code is time, and time is money and money is power | or code is 
 knoweldge and knowledge is power).
 I mean user areas are are a rule to bloated for any conservitive 
 backup.
 Like I have Disc ISO in mine (MDNSAA, linux distros)

 If we wanted to back up user areas it would be a serious undertaking,
 What would be the bandwidth from UCC to Arts?

 I guess Wheel docs, coke logs etc could also do with offsite backup

 Frames
 [*OX]

From atyndall at ucc.asn.au  Fri May 13 19:40:30 2011
From: atyndall at ucc.asn.au (Ash Tyndall)
Date: Fri, 13 May 2011 19:40:30 +0800
Subject: [tech] Fixing the Door Sensors
Message-ID: <BANLkTimOnrXDCTV9bp4xKH8sLxceC7Mv9w@mail.gmail.com>

Because Bob envies my apparent awesome, I was tasked with figuring out
how to make the reed switch detectors work again. Here is my proposal:

Use one of the Arduinos we bought to interface with the reed switch
circuits. The Arduino is also connected to the network. Using the
magic of the Arduino libraries, the Arduino will host an XML or JSON
webpage that has the status of the doors embedded into it. A service
on one of the servers will periodically query the Arduino for this
information, and update the Jabber and finger door at ucc.asn.au stuff
accordingly.

Opinions?

-- 
Ash Tyndall [ASH]
2011 Fresher Rep

From danielax at gmail.com  Tue May 24 18:31:18 2011
From: danielax at gmail.com (Daniel Axtens)
Date: Tue, 24 May 2011 18:31:18 +0800
Subject: [tech] Napoli: Firmware password set
Message-ID: <A29E185E-124E-405C-AB8F-7B9BE6AF5366@gmail.com>

Hi all

Due to the exploits of one of our newer members in entering single user mode on napoli, I have set a firmware password.

The firmware password is the same as the clubroom password.

[DJA]

From atyndall at ucc.asn.au  Wed May 25 13:35:35 2011
From: atyndall at ucc.asn.au (Ash Tyndall)
Date: Wed, 25 May 2011 13:35:35 +0800
Subject: [tech] Club Hosting: VPSs vs Integration
In-Reply-To: <BANLkTikYEnU68HbZ8i39Cr4PaXJFHsLopg@mail.gmail.com>
References: <BANLkTikYEnU68HbZ8i39Cr4PaXJFHsLopg@mail.gmail.com>
Message-ID: <BANLkTimYtd7vwAqVSdG4-fjaH=GWdLsEww@mail.gmail.com>

Recently, the committee approved the provision of UCC services for paid club
hosting. The specifics of the prices and such have already been decided on,
however, there are some implementation issues that I think need discussing.

The initial thought of the committee was that a VPS system would be the
easiest for those clubs which need more than just web hosting services; it
isolates the clubs, allowing them to configure a myriad of services while
still being easily locked in case of non-payment.

However, VPSs come with certain security issues that make them not the most
ideal technology for club hosting, even if that club needs additional
services. The fact that they clubs are in charge of the security of the
server means such things may be neglected.

For example, the UWA Atheist & Skeptic Society, who are (beside CSSC) the
only club interested in this service, only need web hosting, some
procmail/postfix routing and a mailman instance to run on their domain
uass.asn.au.

In this case, configuration files would be in a variety of places on the UCC
servers, making account locking difficult if it needs to be done. VPSs solve
this problem.

However, on IRC, [JCF] has raised an objection to the necessity of VPSs,
even for the use case that UASS has.

It was suggested that due to the infrequency of the request, and the lack of
issues with account locking in the past, that the security benefits of
integration outweigh the advantages of VPS-based service isolation.

"We're probably better off sticking with what we do well rather than
introducing something we're less likely to maintain in the interests of
security."

I am personally for integration; it makes it much easier for the club, as
they don't need to perform software-based maintenance. What are other
people's opinions?

-- 
Ash Tyndall [ASH]
2011 Fresher Rep
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20110525/cd4b9c0f/attachment.htm 

From oxinabox at ucc.asn.au  Wed May 25 14:12:18 2011
From: oxinabox at ucc.asn.au (oxinabox at ucc.asn.au)
Date: Wed, 25 May 2011 14:12:18 +0800
Subject: [tech] Club Hosting: VPSs vs Integration
Message-ID: <a22ed0155ca58f960f2311f0beb3dcc2@secure.ucc.asn.au>

 Unisfa has a account with UCC,
 I can't remember what's run out of it,
 I know there is emailing aliases, and maybe koha?

 Unigames has a similar account, for mailing list and a forwarding list.
 Neither of which have ever been used (we sorted guild Out)

 Frames
 [*OX]

From atyndall at ucc.asn.au  Wed May 25 14:17:37 2011
From: atyndall at ucc.asn.au (Ash Tyndall)
Date: Wed, 25 May 2011 14:17:37 +0800
Subject: [tech] Club Hosting: VPSs vs Integration
In-Reply-To: <a22ed0155ca58f960f2311f0beb3dcc2@secure.ucc.asn.au>
References: <a22ed0155ca58f960f2311f0beb3dcc2@secure.ucc.asn.au>
Message-ID: <BANLkTi=gpsai5h_jmXmjysyzbAgGRgKF=g@mail.gmail.com>

UASS's configuration is slightly more involved, because everything is going
through a completely different domain, and the mail routing needs to
conditionally forward mail to either mailman or Google Apps.

On Wed, May 25, 2011 at 2:12 PM, <oxinabox at ucc.asn.au> wrote:

> Unisfa has a account with UCC,
> I can't remember what's run out of it,
> I know there is emailing aliases, and maybe koha?
>
> Unigames has a similar account, for mailing list and a forwarding list.
> Neither of which have ever been used (we sorted guild Out)
>
> Frames
> [*OX]
>



-- 
Ash Tyndall [ASH]
2011 Fresher Rep
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20110525/e7f6dadb/attachment.htm 

From zanchey at ucc.gu.uwa.edu.au  Wed May 25 22:31:09 2011
From: zanchey at ucc.gu.uwa.edu.au (David Adam)
Date: Wed, 25 May 2011 22:31:09 +0800 (WST)
Subject: [tech] Club Hosting: VPSs vs Integration
In-Reply-To: <BANLkTimYtd7vwAqVSdG4-fjaH=GWdLsEww@mail.gmail.com>
References: <BANLkTikYEnU68HbZ8i39Cr4PaXJFHsLopg@mail.gmail.com>
	<BANLkTimYtd7vwAqVSdG4-fjaH=GWdLsEww@mail.gmail.com>
Message-ID: <alpine.DEB.2.00.1105252154450.7643@motsugo.ucc.gu.uwa.edu.au>

On Wed, 25 May 2011, Ash Tyndall wrote:
> Recently, the committee approved the provision of UCC services for paid club
> hosting. The specifics of the prices and such have already been decided on,
> however, there are some implementation issues that I think need discussing.

We approved it *again*?
http://www.ucc.asn.au/infobase/committee/2007/2007-03-26.txt

> The initial thought of the committee was that a VPS system would be the
> easiest for those clubs which need more than just web hosting services; it
> isolates the clubs, allowing them to configure a myriad of services while
> still being easily locked in case of non-payment.
>  
> However, VPSs come with certain security issues that make them not the most
> ideal technology for club hosting, even if that club needs additional
> services. The fact that they clubs are in charge of the security of the
> server means such things may be neglected.

VPS is a somewhat nebulous term that encompasses a broad range of 
technologies; I assume in this specific case you are talking 
about virtualised or paravirtualised machines (VMs).

> For example, the UWA Atheist & Skeptic Society, who are (beside CSSC) the
> only club interested in this service, only need web hosting, some
> procmail/postfix routing and a mailman instance to run on their domain
> uass.asn.au.
> 
> In this case, configuration files would be in a variety of places on the UCC
> servers, making account locking difficult if it needs to be done. VPSs solve
> this problem.

A questionable assertion :-)

> However, on IRC, [JCF] has raised an objection to the necessity of VPSs,
> even for the use case that UASS has.
> 
> It was suggested that due to the infrequency of the request, and the lack of
> issues with account locking in the past, that the security benefits of
> integration outweigh the advantages of VPS-based service isolation.
> 
> "We're probably better off sticking with what we do well rather than
> introducing something we're less likely to maintain in the interests of
> security."
> 
> I am personally for integration; it makes it much easier for the club, as
> they don't need to perform software-based maintenance. What are other
> people's opinions?

I don't really care. Basically it comes down to how much you are willing 
to do yourself (where "you" is the club in question). If you want to 
configure things without having to involve wheel members and are happy to 
deal with all the issues, go for your life. Differences in quality of 
service are probably negligible.

(There is an analogy for in-tree and out-of-tree Linux drivers, but 
frankly I think we've expended enough words on this as is.)

David Adam
UCC Wheel Member
zanchey at ucc.gu.uwa.edu.au

From ian at mckellar.org  Thu May 26 01:09:05 2011
From: ian at mckellar.org (Ian McKellar)
Date: Wed, 25 May 2011 10:09:05 -0700
Subject: [tech] Club Hosting: VPSs vs Integration
In-Reply-To: <BANLkTi=gpsai5h_jmXmjysyzbAgGRgKF=g@mail.gmail.com>
References: <a22ed0155ca58f960f2311f0beb3dcc2@secure.ucc.asn.au>
	<BANLkTi=gpsai5h_jmXmjysyzbAgGRgKF=g@mail.gmail.com>
Message-ID: <BANLkTimNTP-FYBVtsRL_qa1NAXKJ1sXMCw@mail.gmail.com>

Why don't they just use google apps? Are their mail routing needs that
complicated?

Ian

On Tue, May 24, 2011 at 11:17 PM, Ash Tyndall <atyndall at ucc.asn.au> wrote:
> UASS's configuration is slightly more involved, because everything is going
> through a completely different domain, and the mail routing needs to
> conditionally forward mail to either mailman or Google Apps.
>
> On Wed, May 25, 2011 at 2:12 PM, <oxinabox at ucc.asn.au> wrote:
>>
>> Unisfa has a account with UCC,
>> I can't remember what's run out of it,
>> I know there is emailing aliases, and maybe koha?
>>
>> Unigames has a similar account, for mailing list and a forwarding list.
>> Neither of which have ever been used (we sorted guild Out)
>>
>> Frames
>> [*OX]
>
>
>
> --
> Ash Tyndall [ASH]
> 2011 Fresher Rep
>



-- 
Ian McKellar? <http://ian.mckellar.org/>
ian at mckellar.org: email | jabber | msn
ianloic: flickr | aim | yahoo | skype | linkedin | etc.

From trs80 at ucc.gu.uwa.edu.au  Thu May 26 01:34:33 2011
From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha)
Date: Thu, 26 May 2011 01:34:33 +0800 (WST)
Subject: [tech] Club Hosting: VPSs vs Integration
In-Reply-To: <BANLkTimNTP-FYBVtsRL_qa1NAXKJ1sXMCw@mail.gmail.com>
References: <a22ed0155ca58f960f2311f0beb3dcc2@secure.ucc.asn.au>
	<BANLkTi=gpsai5h_jmXmjysyzbAgGRgKF=g@mail.gmail.com>
	<BANLkTimNTP-FYBVtsRL_qa1NAXKJ1sXMCw@mail.gmail.com>
Message-ID: <alpine.DEB.2.00.1105260131420.21518@motsugo.ucc.gu.uwa.edu.au>

On Wed, 25 May 2011, Ian McKellar wrote:

> Why don't they just use google apps? Are their mail routing needs that
> complicated?

Per IRC, they do, but the free version of Google Apps doesn't do mailing 
lists, and using Google Groups wasn't suitable for some reason.

-- 
# 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 atyndall at ucc.asn.au  Sat May 28 14:03:38 2011
From: atyndall at ucc.asn.au (Ash Tyndall)
Date: Sat, 28 May 2011 14:03:38 +0800
Subject: [tech] Some Steam issues resolved?
Message-ID: <BANLkTi=70_dyR-a82-aQOsSviV7U3CFMVw@mail.gmail.com>

I've been fiddling around with the permissions on the D:\Steam folder on the
four windows machines, and I think I've found a configuration that stops it
needing an admin password to login to Steam for the first time on a
computer. It may have also fixed some other permission problems.

I've documented this configuration at http://wiki.ucc.asn.au/Steam.

The basic gist of it is you allow gumby and Everyone all rights except
the ability to change ownership or change the permissions. Then you made
everything in the directory inherit those permissions, ensuring that the
owner is SYSTEM. The problem seemed to be that Steam choked when trying to
create a player directory in D:\Steam\steamapps, which it only does if you
haven't logged in before.

The configuration still allows wheel and winadmin members to change
permissions, but I haven't fully tested their effects. Try not to run
Steam.exe as Administrator if you don't have to; I'm not sure if Steam will
revert the permissions.

-- 
Ash Tyndall [ASH]
2011 Fresher Rep
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20110528/353eeafc/attachment.htm