Gigabit Ethernet SBus & Gigabit Ethernet Sun 420R/220R

[Earl D. Baugh Jr.]

Two Gigabit ethernet questions....

Did anybody ever make a Gigabit ethernet SBus card that ran across  
twisted pair?
I've done some Googling and see ones that look like fiber based, but  
haven't seen any that look like they support twisted pair... (I've got  
a couple SS10's and was wondering if there is anyway to get them  
talking Gigabit...
(I don't have fiber, just the twisted pair..)

Second, if I were looking for some relatively inexpensive Gigabit  
cards for my Sun 420R/220R, what card should I be looking for?  I'm  
looking for the same as above, twisted pair based, and only need a  
single interface for each machine.  It looks like the Sun GigaSwith  
cards might be the right ones... I see some on EPay that look pretty  
inexpensive (pulls from Ultra 10 boxes).  If anybody on the list has  
any they're willing to part with, let me know...
Otherwise I'll probably go the EPay route.

Thanks.

Earl
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Tom "spot" Callaway]

On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
> Two Gigabit ethernet questions....
>
> Did anybody ever make a Gigabit ethernet SBus card that ran across  
> twisted pair?

Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s (64 bit
bus)? Or is it bridging to PCI to get around that somehow?

~spot
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:

> On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
>> Two Gigabit ethernet questions....
>>
>> Did anybody ever make a Gigabit ethernet SBus card that ran across
>> twisted pair?
>
> Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s (64 bit
> bus)? Or is it bridging to PCI to get around that somehow?

So, basically, you are pointing out that SBUS can do about 2.5 times  
the bandwidth that one would realistically get from GigE?  Why is  
that silly then?
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Tom "spot" Callaway]

On Sat, 2008-04-05 at 23:37 -0400, Joshua Boyd wrote:

> On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:
>
> > On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
> >> Two Gigabit ethernet questions....
> >>
> >> Did anybody ever make a Gigabit ethernet SBus card that ran across
> >> twisted pair?
> >
> > Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s (64 bit
> > bus)? Or is it bridging to PCI to get around that somehow?
>
> So, basically, you are pointing out that SBUS can do about 2.5 times  
> the bandwidth that one would realistically get from GigE?  Why is  
> that silly then?

Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.

~spot
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Steve Sandau]

Tom "spot" Callaway wrote:

> On Sat, 2008-04-05 at 23:37 -0400, Joshua Boyd wrote:
>> On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:
>>
>>> On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
>>>> Two Gigabit ethernet questions....
>>>>
>>>> Did anybody ever make a Gigabit ethernet SBus card that ran across
>>>> twisted pair?
>>> Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s (64 bit
>>> bus)? Or is it bridging to PCI to get around that somehow?
>> So, basically, you are pointing out that SBUS can do about 2.5 times  
>> the bandwidth that one would realistically get from GigE?  Why is  
>> that silly then?
>
> Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.
>

I would hope that gigE would max out at more like 500MB/s, but I have
not worked with gigE and jumbo frames enough to know.

100MB ethernet maxes out at what, 70MB/s or 80MB/s under ideal
conditions, or is that even too optimistic?

In any case, there *are* gigE SBUS cards, I have a fiber one (which I am
not using, hint, hint).

Steve
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Apr 6, 2008, at 5:47 AM, Steve Sandau wrote:

> Tom "spot" Callaway wrote:
>> On Sat, 2008-04-05 at 23:37 -0400, Joshua Boyd wrote:
>>> On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:
>>>
>>>> On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
>>>>> Two Gigabit ethernet questions....
>>>>>
>>>>> Did anybody ever make a Gigabit ethernet SBus card that ran across
>>>>> twisted pair?
>>>> Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s  
>>>> (64 bit
>>>> bus)? Or is it bridging to PCI to get around that somehow?
>>> So, basically, you are pointing out that SBUS can do about 2.5  
>>> times  the bandwidth that one would realistically get from GigE?  
>>> Why is  that silly then?
>> Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.
>
> I would hope that gigE would max out at more like 500MB/s, but I  
> have not worked with gigE and jumbo frames enough to know.

GigE is supposedly capable of going close to 100 MB/s a second, with  
jumbo frames.

> 100MB ethernet maxes out at what, 70MB/s or 80MB/s under ideal  
> conditions, or is that even too optimistic?

Typically one uses MB for megabyte, and Mb for megabit.

FastE is 100Mb, but in average circumstances in the real world it  
only does about 7.5-9MB.

GigE is 1000Mb, and is rumoured to be capable of doing 80MB+, but  
that requires fairly good hardware.

Some people say that it isn't worth upgrading to GigE if you can't  
get 10 times faster than FastE, but I wonder why people think that a  
4-6x upgrade for older machines isn't worthwhile?
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Steve Sandau]

<snip>

>>> Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.
>>
>> I would hope that gigE would max out at more like 500MB/s, but I have
>> not worked with gigE and jumbo frames enough to know.
>
> GigE is supposedly capable of going close to 100 MB/s a second, with
> jumbo frames.
>
>> 100MB ethernet maxes out at what, 70MB/s or 80MB/s under ideal
>> conditions, or is that even too optimistic?
>
> Typically one uses MB for megabyte, and Mb for megabit.
>
> FastE is 100Mb, but in average circumstances in the real world it only
> does about 7.5-9MB.
>
> GigE is 1000Mb, and is rumoured to be capable of doing 80MB+, but that
> requires fairly good hardware.
>
> Some people say that it isn't worth upgrading to GigE if you can't get
> 10 times faster than FastE, but I wonder why people think that a 4-6x
> upgrade for older machines isn't worthwhile?

Ah, I was not paying enough attention to Mb vs MB. I did mean megabits
in my statements, and typed MB without thinking. That was a kind way of
correcting me, though. ;-)

Mostly, I was thinking that you should be able to get the same
percentage of ideal bandwidth from GigE (with jumbo frames) as from
FastE. It sounds like that is the case, but only under the best
circumstances. Correct?

And if SBUS maxes out at 200 megabytes, that is 1.6 gigabits, and a GigE
card should certainly be able to move data across the SBUS as fast as
across the network.

As I mentioned before, I have a fiber GigE SBUS card for COS if anyone
is interested.

Steve
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson]

>From: Steve Sandau <ssandau@...>
>Date: 2008/04/06 Sun AM 08:34:01 CDT
>To: The Rescue List <rescue@...>
>Subject: Re: [rescue] Gigabit Ethernet SBus & Gigabit Ethernet Sun 420R/220R

<snip>

>And if SBUS maxes out at 200 megabytes, that is 1.6 gigabits, and a GigE
>card should certainly be able to move data across the SBUS as fast as
>across the network.

IIRC the docs had some very high CPU requirements to support "high speed" network interfaces - I seem to recall something like one CPU per 100 Mb/sec interface on U1/U2-era machines... Made me wonder where people were deploying those QFE cards...

>As I mentioned before, I have a fiber GigE SBUS card for COS if anyone
>is interested.

I dont' want them, but I am curious what it would take to make use of them - obviouosly cards & fiber cables, but it a switch needed? some other technology? and how are the cards configured - is it like an Ethernet connection to the software, it just travels over different media?

I've (obviously) never tinkered with fiber before, but I'm thinking I might want to get a handle on the technology...

Lionel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson]

>From: "Earl D. Baugh Jr." <earl@...>
>Date: 2008/04/05 Sat PM 10:03:55 CDT
>To: The Rescue List <rescue@...>
>Subject: [rescue] Gigabit Ethernet SBus & Gigabit Ethernet Sun 420R/220R

>Two Gigabit ethernet questions....

<snip>

>Second, if I were looking for some relatively inexpensive Gigabit  
>cards for my Sun 420R/220R, what card should I be looking for?  I'm  
>looking for the same as above, twisted pair based, and only need a  
>single interface for each machine.  It looks like the Sun GigaSwith  
>cards might be the right ones... I see some on EPay that look pretty  
>inexpensive (pulls from Ultra 10 boxes).  If anybody on the list has  
>any they're willing to part with, let me know...
>Otherwise I'll probably go the EPay route.

Sun Gigaswifts look to be about $25 shipped on eBay, but if you don't need to boot over the Gigabit interface, the cards I mentioned a few days ago *should* work, though I've not tried one yet (haven't set up gigabit at home):

<quote>
Hello all,

Looking for a confirmation before I pull the trigger - my research has turned
up that the RealTek RTL8169 Gigabit Ethernet cards work with Solaris 10 (but
NOT OpenBoot PROM, so no boot on net)[0], and there is a retailer selling
refurbs for $11.95 (with $1.95 S/H per networking order)[1] - anyone used
these cards in their PCI UltraSPARC boxes (U5/10 and 30/60)?

Lionel

[0] http://unix.derkeiler.com/Newsgroups/comp.sys.sun.hardware/2006-
08/msg00330.html

[1] http://justdeals.stores.yahoo.net/ga311nar.html (I've bought from them
before, never a problem, so a trusted vendor)
</quote>

The Gigaswift looks good, but it wants/needs a 64 bit PCI slot (not a problem, just an observation) but it should be net bootable...

Lionel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Mark Brown-9]

Lionel Peterson wrote:
>
>
> I dont' want them, but I am curious what it would take to make use of them - obviouosly cards & fiber cables, but it a switch needed? some other technology? and how are the cards configured - is it like an Ethernet connection to the software, it just travels over different media?
>
>  
The minimum config is 2 cross connected fibre gig cards of the same
optical wavelength, cross connected. (sans switch)

I had a private GE interconnect between my E3000 and E4500, and a public
facing GE to my core network from each of those as well.

/M
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Phil Stracchino-3]

Steve Sandau wrote:

> Tom "spot" Callaway wrote:
>> On Sat, 2008-04-05 at 23:37 -0400, Joshua Boyd wrote:
>>> On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:
>>>> On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
>>>>> Two Gigabit ethernet questions....
>>>>>
>>>>> Did anybody ever make a Gigabit ethernet SBus card that ran across
>>>>> twisted pair?
>>>> Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s (64 bit
>>>> bus)? Or is it bridging to PCI to get around that somehow?
>>> So, basically, you are pointing out that SBUS can do about 2.5 times  
>>> the bandwidth that one would realistically get from GigE?  Why is  
>>> that silly then?
>>
>> Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.
>>
>
> I would hope that gigE would max out at more like 500MB/s, but I have
> not worked with gigE and jumbo frames enough to know.
>
> 100MB ethernet maxes out at what, 70MB/s or 80MB/s under ideal
> conditions, or is that even too optimistic?

It's megabits, remember, not megabytes.

I have gotten an honest 97Mbit/s data rate across my local 100Mbit
network.  That's about 12 MB/s.  If GigE can manage 900Mbit/s actual
throughput, that's 112MB/s.  So if SBUS tops out at 200MB/s, it should
take GigE traffic just fine.  If that's 200Mbit/s, then GigE on SBUS
isn't going to work too well.  :)



--
   Phil Stracchino, CDK#2     DoD#299792458     ICBM: 43.5607, -71.355
   alaric@...   alaric@...   phil@...
          Renaissance Man, Unix ronin, Perl hacker, Free Stater
                  It's not the years, it's the mileage.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jonathan Katz-2]

On Apr 6, 2008, at 10:46 AM, Phil Stracchino wrote:
> It's megabits, remember, not megabytes.
>
> I have gotten an honest 97Mbit/s data rate across my local 100Mbit  
> network.  That's about 12 MB/s.  If GigE can manage 900Mbit/s  
> actual throughput, that's 112MB/s.  So if SBUS tops out at 200MB/s,  
> it should take GigE traffic just fine.  If that's 200Mbit/s, then  
> GigE on SBUS isn't going to work too well.  :)

I feel like a broken record. I know I've posted this story 2x before :D

Back in 1997 I was a contractor at Alteon networks. We had Tatung  
Ultra 2 clones with the 336Mhz(?) CPUs in them. They were sbus  
systems. With the sbus gigabit cards using 8k frames (jumbo frames)  
and ttcp we could hard-lock the systems and managed to get 800Mbits  
of throughput. This was back in the Solaris 2.5.1 days with 32bit  
kernels, too, and rev 1.x drivers.

Other caveats at the time were that you wanted at least two CPUs in  
the system because one would be busy managing the network traffic,  
even though it would work fine with one CPU.

With those memories fondly in the back of my brain I'm thinking we  
could easily hit 900Mbits or greater with our more refined OSes and  
drivers (and more typical network traffic than ttcp blasting) and  
still have a very functional system.

-Jon
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Apr 6, 2008, at 11:19 AM, Jonathan Katz wrote:

> On Apr 6, 2008, at 10:46 AM, Phil Stracchino wrote:
>> It's megabits, remember, not megabytes.
>>
>> I have gotten an honest 97Mbit/s data rate across my local 100Mbit  
>> network.  That's about 12 MB/s.  If GigE can manage 900Mbit/s  
>> actual throughput, that's 112MB/s.  So if SBUS tops out at 200MB/
>> s, it should take GigE traffic just fine.  If that's 200Mbit/s,  
>> then GigE on SBUS isn't going to work too well.  :)
>
> I feel like a broken record. I know I've posted this story 2x  
> before :D
>
> Back in 1997 I was a contractor at Alteon networks. We had Tatung  
> Ultra 2 clones with the 336Mhz(?) CPUs in them. They were sbus  
> systems. With the sbus gigabit cards using 8k frames (jumbo frames)  
> and ttcp we could hard-lock the systems and managed to get 800Mbits  
> of throughput. This was back in the Solaris 2.5.1 days with 32bit  
> kernels, too, and rev 1.x drivers.

What did you get in real usage outside of a benchmarking program though?
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jonathan Katz-2]

On Apr 6, 2008, at 2:03 PM, Joshua Boyd wrote:
> On Apr 6, 2008, at 11:19 AM, Jonathan Katz wrote:
>> Back in 1997 I was a contractor at Alteon networks. We had Tatung  
>> Ultra 2 clones with the 336Mhz(?) CPUs in them.
> What did you get in real usage outside of a benchmarking program  
> though?

At the time the major server OSes were NT4, Solaris 2.5.1 and such...  
I think we typically saw 300Mbits sustained.

-Jon
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On Apr 6, 2008, at 10:46 , Phil Stracchino wrote:

> Steve Sandau wrote:
>> Tom "spot" Callaway wrote:
>>> On Sat, 2008-04-05 at 23:37 -0400, Joshua Boyd wrote:
>>>> On Apr 5, 2008, at 11:27 PM, Tom "spot" Callaway wrote:
>>>>> On Sat, 2008-04-05 at 23:03 -0400, Earl D. Baugh Jr. wrote:
>>>>>> Two Gigabit ethernet questions....
>>>>>>
>>>>>> Did anybody ever make a Gigabit ethernet SBus card that ran  
>>>>>> across
>>>>>> twisted pair?
>>>>> Wouldn't that be a bit silly, since SBUS maxes out at 200 MB/s  
>>>>> (64 bit
>>>>> bus)? Or is it bridging to PCI to get around that somehow?
>>>> So, basically, you are pointing out that SBUS can do about 2.5  
>>>> times  the bandwidth that one would realistically get from GigE?  
>>>> Why is  that silly then?
>>>
>>> Ehh, I suppose. I was thinking 1000 MB/s, but that's not realistic.
>>>
>> I would hope that gigE would max out at more like 500MB/s, but I  
>> have not worked with gigE and jumbo frames enough to know.
>> 100MB ethernet maxes out at what, 70MB/s or 80MB/s under ideal  
>> conditions, or is that even too optimistic?
>
> It's megabits, remember, not megabytes.
>
> I have gotten an honest 97Mbit/s data rate across my local 100Mbit  
> network.  That's about 12 MB/s.  If GigE can manage 900Mbit/s actual  
> throughput, that's 112MB/s.  So if SBUS tops out at 200MB/s, it  
> should take GigE traffic just fine.  If that's 200Mbit/s, then GigE  
> on SBUS isn't going to work too well.  :)

On my gigabit LAN, I seem to max out usually around 20MBytes/sec.

That's going from a Mac Pro to a Dell T105.

Neither of them is using enough CPU to even show up during a transfer,  
so I'm not sure why I can't get better.

Of course, my NAS box maxes out at less than that, around 14MBytes/
sec, but it has a tiny 200MHz embedded CPU.

Seems to me CPU isn't the issue, nor should the bus be an issue with  
my Mac and my Dell, both of which have gigE on PCI Express.

My switch is an HP 1800 24-port, which is supposed to have 48Gbit/sec  
of bandwidth.

--
Shannon Hendrix
shannon@...
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Sridhar Ayengar]

Shannon Hendrix wrote:

> On my gigabit LAN, I seem to max out usually around 20MBytes/sec.
>
> That's going from a Mac Pro to a Dell T105.
>
> Neither of them is using enough CPU to even show up during a transfer,
> so I'm not sure why I can't get better.
>
> Of course, my NAS box maxes out at less than that, around 14MBytes/sec,
> but it has a tiny 200MHz embedded CPU.
>
> Seems to me CPU isn't the issue, nor should the bus be an issue with my
> Mac and my Dell, both of which have gigE on PCI Express.
>
> My switch is an HP 1800 24-port, which is supposed to have 48Gbit/sec of
> bandwidth.

I get the same kind of numbers on the same model switch.  Maybe the
switch just sucks.  I should check if there's a firmware upgrade
available for it.

Peace...  Sridhar
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Michael-John Turner-2]

On Fri, Apr 25, 2008 at 12:02:21AM -0400, Shannon Hendrix wrote:
> Neither of them is using enough CPU to even show up during a transfer, so
> I'm not sure why I can't get better.

You should be able to get better performance if you enable jumbo frames -
doing so does require all machines connected to the switch or on the same
VLAN to have them enabled though.

-mj
--
Michael-John Turner
mj@...     | http://mjturner.net/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson]

>From: Shannon Hendrix <shannon@...>
>Date: 2008/04/24 Thu PM 11:02:21 CDT
>To: The Rescue List <rescue@...>
>Subject: Re: [rescue] Gigabit Ethernet SBus & Gigabit Ethernet Sun 420R/220R

<snip>

>On my gigabit LAN, I seem to max out usually around 20MBytes/sec.
>
>That's going from a Mac Pro to a Dell T105.
>
>Neither of them is using enough CPU to even show up during a transfer,  
>so I'm not sure why I can't get better.
>
>Of course, my NAS box maxes out at less than that, around 14MBytes/
>sec, but it has a tiny 200MHz embedded CPU.
>
>Seems to me CPU isn't the issue, nor should the bus be an issue with  
>my Mac and my Dell, both of which have gigE on PCI Express.
>
>My switch is an HP 1800 24-port, which is supposed to have 48Gbit/sec  
>of bandwidth.

Cabling? Have you tried a direct gigabit (crossover) connection? Also, your T105 is a "budget" system, which can cost less as a complete system that a good, server-grade Intel (or other) gigabit Ethernet adapter... (well, I exagerate the price difference, but the the point is valid I think, the on-board gigabit Ethernet adapter is not the best available, and better adapters do make a difference.)

Lionel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[KRM]

Jumbo frames can make a huge difference, especially on "bus impaired" older, low end stuff.

I went to great lengths to keep a large portion of the network at my old job jumbo clean.  I just setup a small linux box to route between the clean and unclean subnets.  There will always be some stuff, like JetDirect cards, that are not compliant so some type of go between will be necessary.

The main annoyance in doing so is finding out if a particular NIC is actually jumbo clean or not and if so, to what degree.  Not all jumbo clean devices support MTUs all the way up to 9000.  Getting this info from the manufacturers can often be a difficulty.

K



On Fri, Apr 25, 2008 at 09:35:21AM +0100, Michael-John Turner wrote:

> On Fri, Apr 25, 2008 at 12:02:21AM -0400, Shannon Hendrix wrote:
> > Neither of them is using enough CPU to even show up during a transfer, so
> > I'm not sure why I can't get better.
>
> You should be able to get better performance if you enable jumbo frames -
> doing so does require all machines connected to the switch or on the same
> VLAN to have them enabled though.
>
> -mj
> --
> Michael-John Turner
> mj@...     | http://mjturner.net/
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Dan Sikorski]

Sridhar Ayengar wrote:
>
> I get the same kind of numbers on the same model switch.  Maybe the
> switch just sucks.  I should check if there's a firmware upgrade
> available for it.

On an almost-related note, at some point in the past, I had discussed on
this list (with Sridhar?) the possibility of using a WS-X5410 9 port
gigabit module with WS-G5483 1000baseT GBICs in a cisco catalyst  5000.  
Even though cisco documentation says that the copper GBICs will not
work, they do work just fine.  CatOS sees all ports, and shows the ports
with the copper GBICS as 1000baseT.

There are a couple of limitations in my setup, however.  Since the
catalyst 5000 only has a 1.3Gbps backplane, performance between the FDDI
and 10/100 ethernet portions of the network could be limited.  This is
not a problem for me since we're talking about my home network with only
a handful of hosts.  Due to an ASIC limitation, the WS-X5410 cannot use
jumbo frames, which sucks.  Had i known that, i probably would have
opted for getting a couple of the 3 port modules, and living with the
backplane limitation between the two modules, or finding  a 5500 chassis
to move everything to.

    -Dan Sikorski
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue