Next Meeting - 24th February

[Duncan Roe]

Hi everyone,

The next meeting is 4 weeks from yesterday.  I reset the attendance list
today and have just updated https://mlug-au.org.  Here's the URL for the
online meeting anyway: https://meet.jit.si/mlug-au

We have yet to schedule any talks: does anyone have something they'd
like to share?

Do put your name on the attendance list (on the website) if you'd like
to have an in-person meeting again next month. We only need 4 names,
mine's there already,

Cheers ... Duncan.

[Duncan Roe]

Hi everyone,

Next meeting is now 1 week from today.

We only have 3 names on the attendance list:

I'll attempt to book the GNH when we get a 4th.

As well as Kevin's talk on Cloudburst,

I could give a lightning talk on progress I've made with per-xterm bash histories

if that would interest anyone.

Cheers ... Duncan.

[Andrew McGlashan]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi,

On 17/2/25 4:05 pm, Duncan Roe wrote:
> Next meeting is now 1 week from today.
>
> We only have 3 names on the attendance list: I'll attempt to book the GNH when we get a 4th.

I'm not sure I can make it, but it may be possible for me (depends on a birthday and the arrangements).

> As well as Kevin's talk on Cloudburst,

Cloudburst.... a lot different to Syncthing? Is Syncthing something that people have any issues with?
- https://syncthing.net/

I refuse to use big tech as much as possible; especially, but not limited to the current AI offerings.

Found fairly recently when I installed the latest MS Office for a client that they were pushing a "new" Outlook that
pushes all emails to the cloud! Had to find original Outlook for classic behaviour and that will likely stop working in
or around 2029 I think (if I remember correctly). Of course I would prefer the client used Thunderbird and also Linux
instead of Winblows for that matter, but sadly not an option for them.

> I could give a lightning talk on progress I've made with per-xterm bash histories if that would interest anyone.

Yes, that would interest me.

I have an issue with XFCE terminals, I run about 9 terminals with a number of tabs for each; each terminal window has it's
own "task related theme" of tabs. I had toyed with the idea of launching all the windows (XFCE terminal windows) via a
script, but didn't find a way to do it well and reliably -- besides some of the terminal windows operate screen sessions
local and remote, some run as root user with most running as a standard user, so it can be a bit of a pain to set them all
up after a reboot; I copy and paste from a "helper" text file, adjusting screen titles as well.

CLI bash histories, unique to each tab would be nice too. I regularly put in a space before commands I don't want in
history and re-use recent history in the tab easily with simple up arrows. Not sure distinct histories would be that
useful, but it might be.

> Cheers ... Duncan.
>
> On Tuesday, January 28, 2025 at 8:32:58 PM UTC+11 Duncan Roe wrote:
>
> Hi everyone,
>
> The next meeting is 4 weeks from yesterday. I reset the attendance list today and have just updated
> https://mlug-au.org. Here's the URL for the online meeting anyway: https://meet.jit.si/mlug-au
>
> We have yet to schedule any talks: does anyone have something they'd like to share?
>
> Do put your name on the attendance list (on the website) if you'd like to have an in-person meeting again next month.
> We only need 4 names, mine's there already,
>
> Cheers ... Duncan.

Cheers
AndrewM
-----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCZ7LSAwAKCRCoFmvLt+/i
+0pmAP96c0wjdnZ68vNsGpl3Sqe47S0+TqakVTFwDOnPfbWKGwD+KO+3++erGp2d
CHLCddwvDYUDbMDQ6oYF3z9givTRMGE=
=71oF
-----END PGP SIGNATURE-----

[Kevin Exton]

> > As well as Kevin's talk on Cloudburst,
>
> >> Cloudburst.... a lot different to Syncthing?  Is Syncthing
> something that people have any issues with?
>  - https://syncthing.net/

hehe... Cloudbus. No, not like Syncthing. The dream is to be a framework
that lets people build something like Syncthing easier without getting
bogged down in what the best way to route application data streams
across a network should be.

More to say next week. But the general gist of the issue is that people
start to have really poor intuitions about what the best application
layer routing decisions will be once the number of parallel nodes climbs
up over 5 or so. This creates all sorts of shenanigans when it comes to
performance and reliability, and it has let many enterprise frameworks
get away with complete nonsense.

-- Kevin

[Andrew McGlashan]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi,

On 17/2/25 5:24 pm, Kevin Exton wrote:
>>> As well as Kevin's talk on Cloudburst,
>>
>>>> Cloudburst.... a lot different to Syncthing? Is Syncthing something that people have any issues with?
>> - https://syncthing.net/
>
> hehe... Cloudbus. No, not like Syncthing. The dream is to be a framework that lets people build something like
> Syncthing easier without getting bogged down in what the best way to route application data streams across a network
> should be.

Okay, great and yes I noticed my error Cloudburst / Cloudbus ... after pressing send, sorry about that.

> More to say next week. But the general gist of the issue is that people start to have really poor intuitions about what
> the best application layer routing decisions will be once the number of parallel nodes climbs up over 5 or so. This
> creates all sorts of shenanigans when it comes to performance and reliability, and it has let many enterprise
> frameworks get away with complete nonsense.

Sounds good. If you want to send me something off list, then I may include someone else whom might be interested in
attending, also from a long time back. She has been working on something for a long time that may tie in or not; if you
are happy to send me a link to something or have more details then I'll follow up with her.

btw I was thinking it was March already, so that frees up the 24th for me... it's been a long time since I've been to an
MLUG meeting.

Cheers
A.

-----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCZ7LY6QAKCRCoFmvLt+/i
+5jYAP91Tmn0g3rmzNOeAa1CEYM7FOWLFyYkzCTsejrB5aBu+wEA1rlzGAmZzKKQ
L7sfmfm82YNFbI/gGWSwPIqFr6Xc98Q=
=v9Ya
-----END PGP SIGNATURE-----

[Andrew McGlashan]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi,

On 17/2/25 5:24 pm, Kevin Exton wrote:

> More to say next week. But the general gist of the issue is that people start to have really poor intuitions about what
> the best application layer routing decisions will be once the number of parallel nodes climbs up over 5 or so. This
> creates all sorts of shenanigans when it comes to performance and reliability, and it has let many enterprise
> frameworks get away with complete nonsense.

The other thing that sprang to mind was IPFS ...
- https://ipfs.tech/

A.
-----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCZ7LbOQAKCRCoFmvLt+/i
+9+lAQCXISCdqZ/84Pfm4doFOSMQnM/pRmq+XIY8au3NDyD50gEAswl9lFJqk+bm
Dw4aLaBEjLD8nEOoRWF2L8cs0DqnQPY=
=dw5b
-----END PGP SIGNATURE-----

[zak martell]

I would personally suggest staying away from ipfs! I moved to it about 2-3 years ago, and am really struggling to move off of it. It's nothing but problems. I barely store like 3TB/50k files in it, and it constantly stops responding to http/ipfs requests, and always requires a full service restart.

I have a mix of 3-4 physical sites of about 60TB+(per site..) and have/had mix of ipfs, garagehq (https://garagehq.deuxfleurs.fr), seaweedfs, minio, simple rsync copying (with SMB share), so am very curious about this presentation. I even considered https://webtorrent.io 

See you all next week :) 

--
You received this message because you are subscribed to the Google Groups "mlug-au" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mlug-au+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mlug-au/fc626d84-17a2-a0c0-931f-1a0f7af4b464%40affinityvision.com.au.

[Kevin Exton]

> More to say next week. But the general gist of the issue is that people start to have really poor intuitions about what
> the best application layer routing decisions will be once the number of parallel nodes climbs up over 5 or so. This
> creates all sorts of shenanigans when it comes to performance and reliability, and it has let many enterprise
> frameworks get away with complete nonsense.

>> Sounds good.  If you want to send me something off list, then I may include someone else whom might be
>> interested in attending, also from a long time back.  She has been working on something for a long time that may
>> tie in or not; if you are happy to send me a link to something or have more details then I'll follow up with her.

Hi Andrew,

I have a repository: https://github.com/kcexn/cloudbus

But this project is seriously so fresh that I don't have much more than what I will be presenting on the 24'th. I've attached the slides I'll be presenting. The cliff notes are:

• Clever request routing can dramatically reduce request response times and improve application availability when it comes to distributed services.
• Application layer load-balancers don't implement anything near as clever as they should, and transport layer load-balancers shouldn't rely on having any visibility of application state.
• Cloudbus is a session-layer solution aimed at creating a protocol that will be similar to (but not the same as) SIP. Cloudbus will have less telco related overheads that I think are probably related to IP issues (like patents).
  
• The priority is to create something that is easy to build on and extend, without the need for enormous amounts of standardized infrastructure. SIP, for instance, has too many things that you MUST implement in order to have a compliant SIP stack. That's fine if you're a massive company, but not manageable if you are working in a small team, or as in my case, were a researcher.
• If the dream is realized, Cloudbus will make building robust distributed applications easy, and lets application developers delegate the important routing decisions to Cloudbus administrators.
• Right now, I have minimum working examples of the Cloudbus components and am looking for people who might like to play with experimental software so I can start to get a better sense of what kinds of traffic profiles to support.
  

It's probably better just to come along and listen to my presentation, we can chat then. I'm not trying to build Rome in one day.

Best,
Kevin

[Kevin Exton]

> > More to say next week. But the general gist of the issue is that
> people start to have really poor intuitions about what
> > the best application layer routing decisions will be once the number
> of parallel nodes climbs up over 5 or so. This
> > creates all sorts of shenanigans when it comes to performance and
> reliability, and it has let many enterprise
> > frameworks get away with complete nonsense.
>
> >> The other thing that sprang to mind was IPFS ...
> >>  - https://ipfs.tech/

Never thought about running a filesystem work over Cloudbus. But if
there's interest, I'm open to making it work.

I'm literally calling it Cloudbus because you reason about network
services as if it is physically patched into a bus. There are bus
segments, and a bus controller, and the controller coordinates traffic
to whatever network services you need to make your application work
(which in the bare minimum case, is just one, making Cloudbus just a
simple network proxy).

It doesn't however, literally work like a bus, because modern networks
have more features available to them than most simple hardware buses.

-- Kevin

[Duncan Roe]

Hi Everyone,

Do we want to meet in person this month?

There are still only 3 names on the list - is everyone else planning to stay at home

and join online?

With only 3 names on the list, I haven't tried to book the GNH yet.

Remember, the list is only if you *plan* to come - of course you may not manage

on the day owing to work, family &c.

Zak, you said you posted why you don't put your name but I never saw that bit -

could you post it again?

Cheers ... Duncan.

[Duncan Roe]

Hi everyone,

The next meeting is tomorrow 24th Feb.

If all 7 on the list can make it, that will be a post-Covid revord.

See you tomorrow,

Cheers ... Duncan.

On Tuesday, January 28, 2025 at 8:32:58 PM UTC+11 Duncan Roe wrote:

[Andrew McGlashan]

Hi,

On 23/2/25 8:29 pm, Duncan Roe wrote:
> The next meeting is tomorrow 24th Feb.
> If all 7 on the list can make it, that will be a post-Covid revord.

It's not really post-COVID ....

https://www.health.gov.au/topics/covid-19/reporting

It is just not reported daily as it was, but there are still new ICU cases and variants doing the rounds.

Cheers
A.

[Malcolm Herbert]

Folks - I'll be an apology for tonight at the GNH, but I might be able to make it online, NBN permitting ...

Regards,
Malcolm

--
Malcolm Herbert
mj...@mjch.net

[Duncan Roe]

Hi Andrew,

On Sun, Feb 23, 2025 at 11:02:31PM +1100, mlug wrote:
> Hi,

>
> It's not really post-COVID ....
>

My bad. I *meant* "a record attendance since the onset of Covid". Anyway, we're
down to 6 now, unless others (you?) turn up unannounced.

Cheers ... Duncan.

[Duncan Roe]

Hi Andrew,

On Mon, Feb 17, 2025 at 05:07:08PM +1100, mlug wrote:
> Hi,
>
> On 17/2/25 4:05 pm, Duncan Roe wrote:
> >
> > I could give a lightning talk on progress I've made with per-xterm bash
> > histories if that would interest anyone.
>
> Yes, that would interest me.
>
> I have an issue with XFCE terminals, I run about 9 terminals with a
> number of tabs for each; each terminal window has it's own "task related
> theme" of tabs.

I think per-screen histories would be ideal there.

> I had toyed with the idea of launching all the windows
> (XFCE terminal windows) via a script, but didn't find a way to do it
> well and reliably -- besides some of the terminal windows operate screen
> sessions local and remote, some run as root user with most running as a
> standard user, so it can be a bit of a pain to set them all up after a
> reboot; I copy and paste from a "helper" text file, adjusting screen
> titles as well.

If you can make it tonight (online or in person) we could see how far we can get
with XFCE. My examples use FVWM.

>
> CLI bash histories, unique to each tab would be nice too. I regularly
> put in a space before commands I don't want in history and re-use recent
> history in the tab easily with simple up arrows. Not sure distinct
> histories would be that useful, but it might be.
>

I use Control-R[1] and a few shopts that I'll put in the slides now you mention
this topic.

Cheers ... Duncan.

[1] man bash /READLINE /Searching

[Ben Blasco]

Hi all,

Nice to meet you all. I should be able to attend remotely from 7:30 onwards this evening as a first time attendee. I'll give Jitsi a crack.

By way of introduction, I am a Linux enthusiast who works for Red Hat here in Melbourne as a pre-sale Solution Architect, but want to be clear that my interest in attending is as an enthusiast wanting to broaden my knowledge and contacts, and not as a Red Hatter with something to sell.

Hope to see you online later this evening!

Cheers,

Ben

--
You received this message because you are subscribed to the Google Groups "mlug-au" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mlug-au+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/mlug-au/Z7u%2B5JNdStsvzAvT%40slk15.local.net.

[Andrew McGlashan]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi Duncan,

Of course, it wasn't entirely without a bit of tongue in cheek, but it is still a serious issue for some.

Cheers
A.
-----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCZ7wVVwAKCRCoFmvLt+/i
+0DdAQCQHDnKM4wpYu4czA3ftJpfXGWgthElX3P4nd4j4Re2ngD/fwZFjl//WeQy
AzqteEueSqQC9R63BhB/k01NHR0w1v8=
=TQx0
-----END PGP SIGNATURE-----

[dunca...@optusnet.com.au]

Hi Ben,

Great!

Cheers ... Duncan.

----- Original Message -----

From:

mlu...@googlegroups.com

To:

<mlu...@googlegroups.com>

Cc:

Sent:

Mon, 24 Feb 2025 17:06:56 +1100

Subject:

Re: [MLUG] REMINDER: Next Meeting - 24th February

To view this discussion visit https://groups.google.com/d/msgid/mlug-au/CACf%2BMuNzyDr0c947mWGhywmT7MnF%2Bg3JgJzW1-XtVF7myUigyg%40mail.gmail.com.

---

Email sent using Optus Webmail