Re: [9fans] venti/mirrorarenas usage

[Noam Preil]

That's familiar enough that TBH I probably saw your message and forgot
to respond. I was going through Some Shit at the time.

I'm cleaning up my inbox today; I'll respond if I see it and reach out
if I don't :)

- Noam Preil

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M5131c346f1f8a2251fcfa411
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Where can I find active Plan 9 communities for support and collaboration?

[noam]

Quoth kalona.ayeli...@fastmail.us:
> I didn't get it wrong, I was asked to send money if I wanted help, which I 
> found unusual.

If you're sincerely trying to be helpful, then I apologize for
the excessive snark, but frankly I assumed that you were being
obtuse intentionally, since you have repeatedly ignored people
explaining why your posts were not being received well at all.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T9804faa2e50a80d8-M2f89e710cb1448a7514124e0
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth wb.kl...@gmail.com:
> Noam is right in most of his text.

Well of course, I'm me, what do you expect? :D

> But I have to add that the following sentence should be taken with some grain 
> of salt.

Everything should be taken with at least one grain of salt, preferably
more; salt is essential to taste!

> > If the index is on a separate drive, though - e.g. index on SSD, data on 
> > HDDs - mirrorarenas can be used to keep the arenas in sync between multiple 
> > (sets of) HDDs, so that if one fails, venti can use the working drive while 
> > continuing to use the index on the SSD.
> 
> You are better off putting the arenas on SSD, not the index, if you don't 
> want to stress your SSD. Of course, index on SSD looks good at 1st sight for 
> performance, but the  buckets have to be rewritten and reordered at each 
> addition .

... venti's entire index sync logic is horrifying, to be quite honest. 

> 
> If you want the performance gain, and have enough main memory , you should 
> better try my mventi (which uses standard arenas and no index file). -> see 
> other thread

i'm unsure which thread you're talking about; can you link me to
more info on mventi? I've been working on a better venti implementation
as well [1], and it'd be nice to have another reference :)

[1] https://git.sr.ht/~pixelherodev/neoventi


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M85e24fbe45432b086eddb924
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth kalona.ayeli...@fastmail.us:
> While I've received some help, it wasn't clear. I haven't asked for further 
> clarification because I know I would receive more banter for asking simple 
> follow-up questions. 

If you ask questions, I'll happily answer them.

Banter you've received was specifically because you WERE NOT asking questions.



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M13ff4acdb98fc45dcc318530
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth kalona.ayeli...@fastmail.us:

> I see this as a documentation problem. If people can't find information 
> easily, they ask. Without credible documentation, the cycle never ends. We 
> have endless discussions on 9fans on how to do things. Improving 
> documentation is a step forward. Is asking to improve documentation 
> unreasonable? Clearly you don't want me to write it. 

It's a reasonable desire, and one I agree with!

But LLM output is _not_ documentation, and is _not_ improvement.

The people who need the documentation the most will not know enough
to realize just how horribly inaccurate all of the misinformation you
have provided is.

> I am attempting to be helpful.  Even for someone as stupid as I am, I see 
> Plan 9 lacks clear documentation. 

This is a legitimate problem.

Generating incorrect fake documentation will only make this worse.



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M7797be1edbd475ec288f34fd
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Where can I find active Plan 9 communities for support and collaboration?

[noam]

Quoth kalona.ayeli...@fastmail.us:
> I didn't get it wrong, I was asked to send money if I wanted help, which I 
> found unusual.

No, I sarcastically said that if I were to continue reviewing your
LLM-generated spam, I'd start charging for it. At no point did you
give any signs of wanting actual help, which I and many others would
gladly provide free of charge.

> I am using the Plan 9 wiki (https://9p.io/wiki/plan9/plan_9_wiki/) for 
> information. I've read the manual and docs there, but the documentation is 
> lacking. I'm not sure how often the docs, manuals, and ISO file are updated. 
> 
> I am not the only one with Plan 9 issues. Here is another person with similar 
> problems: https://driusan.github.io/plan9.html. 
> 
> There just isn't a consistent community experience here. Why can't newcomers 
> just go to the Plan 9 website, https://9p.io/wiki/plan9/plan_9_wiki/, and get 
> everything they need? What is up with the community site and 9fans?

These are valid points.

Lack of good documentation, especially for a total newcomer or
someone who isn't already technically inclined, is a serious issue.

- Noam Preil



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T9804faa2e50a80d8-M6adccb6249dc608b4e170cdb
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Where can I find active Plan 9 communities for support and collaboration?

[noam]

Quoth kalona.ayeli...@fastmail.us:
> If you're struggling, show your work so others can see and comment on your 
> mistakes

Ironic to hear that from someone who is clearly using LLMs and avoiding
putting in any effort, and simultaneously refusing to show the minimal
work involved (the prompt) while also attempting to illegally claim copyright
on the output :P



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T9804faa2e50a80d8-Mb8f33c281c4be320e469d723
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Where can I find active Plan 9 communities for support and collaboration?

[noam]

Quoth kalona.ayeli...@fastmail.us:
> I'm not a troll; frustrated, perhaps. I just don't understand this culture. I 
> don't think the issue lies with LLMs, but rather with an unwillingness to 
> cooperate unless there's payment involved.

To be totally clear, I was being sarcastic. I thought that was obvious.

> an unspoken social contract and various hidden challenges.

You were told _explicitly_ at least five separate times that the way
you were treating other people was disrespectful, and that the usage
of LLMs was considered as spam by the community.

It's genuinely hard to believe that you aren't a troll at this point.

> I just don't understand this culture. I'm simply seeking a friendly Plan 9 
> group where people genuinely enjoy helping newcomers.  
> I see some people get treated well and others are treated poorly.  If someone 
> asks a simple question, you give the person a pass. If someone asks a hard 
> question, then you treat the person badly. The lesson there is "Don't ask 
> hard questions." That is how I view it. If there are better Plan 9 groups out 
> there, then I'd like to join that group. 

I'm personally happy to answer any questions you want to ask,
simple or hard. The only requirement is that you not use an LLM
to generate them, and that you not deliberately give bad answers
to other people's questions.

It's really not a high bar to clear.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T9804faa2e50a80d8-Mea5814a8454396622e8cb092
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth kalona.ayeli...@fastmail.us:
> khm, you can think whatever you like. If true, then all you can do is let it 
> be.
> 
> My point is accurate information should be easy to find and read

if you care about accurate information, the most helpful thing you can do
is stop using LLMs.

Sincerely, please, please stop. Every single part of what you have posted
has been wrong, and when corrected, you respond by posting even more
misinformation.

Lying to people is not helpful.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M4907d1369937f967aadc5344
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth kalona.ayeli...@fastmail.us:
> Would creating standard operating procedures for newcomers be beneficial? 
> Having procedures detailing how to administer Venti mirroring on a Plan 9 
> system in a wiki seems reasonable. Here is a formatted example of an SOP. I 
> am sure something better could be created.

I understand you're trying to be helpful, but giving people incorrect
information _actively hurts_ people.

Please, please stop.

- Noam Preil



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M11a5d0422f684797644b99f6
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] venti/mirrorarenas usage

[noam]

Quoth Marco Feichtinger :
> venti/mirrorarenas is undocumented, and I couldn't find any topic here,
> which goes into more detail.
> 
> So I am curious how does it work, 
> how does one to set it up, so the arenas get mirrored automatically,
> and why do you use it instead of fs(3) mirror?

fs(3) can be used to mirror a _block device_. In principle, you can use this to 
mirror a full venti - or the arenas! - but you can also use it to mirror 
anything else, it doesn't care about the format of the data.

venti/mirrorarenas should work to _update_ a mirror of an arenas partition: it 
mirrors _just_ the data log of the venti, and provides incremental backup. if 
you want it to be automatic, that's just a cron job; there's nothing fancy for 
it in venti itself.

An important note is that, since it _only_ mirrors the arenas themselves, the 
destination is not a usable venti - for that, you'd need to build the index. In 
principle, if you keep the arena _separate_ from the indexes - i.e. on a 
different drive - this should allow for simply swapping which copy of the 
arenas is used _without updating the index_.

There's also wrarena+rdarena. That needs to be used between _two instances of 
venti_ - i.e. the destination needs a full venti server, not just a disk - but 
has the advantage of allowing the data sets to not be _competely_ identical, 
and since it goes through the server, it ensures the index is up to date on the 
receiving end.

There's a lot of different options on how to keep a venti system replicated / 
backed up, depending on the goal. For just backups, on one machine, where the 
entire venti is on one drive, fs(3) is probably the easiest approach - it can 
be configured to mirror venti across multiple drives, and the bootup scripts 
can automatically fall back to the mirror if the first drive fails.

If the index is on a separate drive, though - e.g. index on SSD, data on HDDs - 
mirrorarenas can be used to keep the arenas in sync between multiple (sets of) 
HDDs, so that if one fails, venti can use the working drive while continuing to 
use the index on the SSD.

Alternatively, for backing up _across the network_, scripts based on 
wrarena+rdarena are probably the way to go. This allows for mirroring between 
multiple separate ventis, even if the configuraiton of the arenas and indices 
are completely different.

What to use and how to set it up depends, as always, on what you're trying to 
do :)

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-Mfca7d8f04716a17aa3283cc6
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

I... this is still completely wrong. I'm not going to bother correcting
you a third time; if you want that, I'm going to have to start charging.
My PayPal should be the same as the email I'm sending this from; feel
free to drop $20 in there if you want me to review your, um, "work".

For what it's worth, I saw this thread and initially intended to provide
a meaningful answer; I had no energy left to spend on email after
dealing with the spam.

I'll try to write up an actually helpful answer today.

Anyone who has not decided that using up water in areas where fresh
water is rationed, as well as burning through multiple kilowatt-hours of
electricity, is a suitable replacement to just executing festoon(1), is
welcome to remind me later if I forget to <3

Lesson learned: next time, I'm only responding to humans, not bots.
At this rate, I'm going to toss fastmail in my upas inbox rejection
script.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-Mef3b91c0ac8c2944475f8a04
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

Other than the fact where that isn't copyrightable?

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M10015b3eb2304befe446229f
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

If someone is hungry and you try to help them by knocking them
unconscious, no, I will not respect your attempt to help.

Good intentions, backed up by incompetence and a lack of respect for
those you seek to help, do not justify harmful actions.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-Meda55a56079d16aee4295b41
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

Are you honestly claiming that you ran the scripts in your post on that
hardware?


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M50d77e27d59b479a0357bb67
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

This is... strictly better, I suppose. New list of mistakes:

- The usage of venti/conf is incorrect. Frankly, I don't know why you'd
even be showing instructions for configuring a venti when the question
was about mirroring capabilities for an existing venti, as I understand
it? It's both irrelevant and incorrect.
- That configuration remains invalid, even if you happen to have both a
SATA and NVMe drive. I gave those two drive paths as examples, if you
just copy them you're _going_ to lose data lol.
- The `index` command actually NAMES the index. This is, quite
honestly, incredible stupid on venti's part, but you actually need a set
of `isect /path/to/indexsection` directives, then an `index [name]`
directive which creates an index across those sections.
- Also, the directive listing an arenas partition is `arenas`, _not_
`arena`, because venti does not use the partition as _an_ arena, but as
a large file in which to store _many_ arena_s_.
- Also also, the _design_ of the venti system is just bad. It makes no
sense to put the index on a SATA drive when you have NVMe available in
basically any scenario.
- Assuming an NVMe drive, as in the MNT Reform, which has _already been
partitioned manually_, a reasonable minimal config can look like this,
off the top of my head:

index main
isect /dev/sdN0/index
arenas /dev/sdN0/arenas

but this is not helpful, because giving you that config without telling
you how to partition or format the partitions is as good as sabotage,
since you're _going_ to make a mistake and lose literally all of your
data. 

- The formatting commands would create an illegal venti, that would fail
to start up.

> venti/fmtarenas -a 512M arena /dev/sdN0/arena
> venti/fmtarenas -a 512M arena /dev/sdE0/arena

You cannot give both arenas partitions the "arena" name, that makes no
sense.

- The mirroring script is incorrect. It's mirroring the data from the
first SATA drive on to two other SATA drives, _but is not_ mirroring the
data from the NVMe drive at all!

- Morever, the comparison to fs(3) remains entirely incoherent, and
demonstrates a total lack of understanding of both venti and fs(3).

- There is no sufficient distinction made between "redundancy" and
"backup/restoration purposes," such that you're recommending two
distinct modes of backup with no proper argument made for when each one
should be used.

- Worse, wrarena and rdarena are used incorrectly. They don't do what
the spambot thinks they do.

- You're attempting to assert copyright over spambot-generated
"content," which is legally not copyrightable. You do not have any legal
rights over that content, and thus cannot reserve any.

- You have demonstrated a lack of understanding of venti. You have
demonstrated a lack of understanding of Plan 9. Worst, by continuing to
post AI spam after being informed of community standards, you've
demonstrated a lack of respect for this community.

You're wasting other people's time, because other people aren't just
using spam generators to reply to yours. Please stop immediately.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-Mc54f442557ee47ffaea970c5
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

Every single part of this is wrong.

- You generally don't want the configuration in a venti.conf file. You
want it inside of one of the venti partitions, which have space reserved
for it. The venti/conf tool is good for this.
- venti configuration files have no `root` command.
- they also have no `log` command.
- They also _REQUIRE_ an index!
- and, generally, an address on which to listen :P
- arenas are not files in /srv, they're partitions, typically e.g.
/dev/sdE0/arena, or /dev/sdN0/arena, or some such, on a SATA or NVMe
device, or flash drive, or SD card, or... you know, some kind of real
storage.
- venti has no internal capacity for mirroring between arenas.
- venti/fmt is not a command. In fact, the need to manually format each
partition of a venti system manually is documented as a bug in the venti
manual page! While venti/fmtindex does point to the config file, it
cannot be run until after the index sections have been manually
formatted!
- `venti` is not the command to run venti; `venti/venti` is.
- `for block in` is not legal rc syntax; that script is not legitimate.
- You can't just... what? No. Just, no. You cannot manually copy a block
from one arena to another by treating arenas as folders, that's
incoherent. Partitions are FILES. The OS / disk drivers have no
conception of what data venti is storing within!

This is very blatantly generated by a spambot. That behavior is not
acceptable within this community. Please refrain from doing this again.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M2fe861959a3bd0e6994cc7b5
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Re: venti/mirrorarenas usage

[Noam Preil]

Legally, you don't own that.

It's already been ruled that, in the US at least, AI spam is not
copyrightable.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tca0eb0fbb2404e31-M596679625e3833139fe94e63
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] List of companies that use Plan 9.

[Noam Preil]

I would _love_ a complete list of known issues.

What I'm aware of:

- IIRC ORCLOSE doesn't work properly (visible as acme leaving temporary
files around after exiting)
- At least one deadlock remaining
- Syncs are not atomic, so a crash can render the system unbootable if
e.g. termrc was being written to disk. I've seen this repeatedly.
- It's not just non-atomic; the resulting state is usually an empty
file. Conjecture: blocks are being allocated, the file points at the new
blocks, the old data has not been copied yet.
- Multiple bugs in the file system checker. Nothing that messes with
data integrity, thankfully! I have needed multiple passes though, as a
single run does not fully clean up the file system. (Noticed this one
when cleaning up after the go failures :P)
- Also seen a deadlock while running `check fix` in the console while
loading the file system, I think, but I don't remember for sure.

I've almost certainly run into more than this. I run into fossil bugs
probably once every two or three months and usually haven't bothered
noting them down because I don't lose data when they happen and they're
rare enough I haven't bothered investigating most of them yet.



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tad3dc0c93039a7d2-Mb1f0845fe06285e7ec50c554
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] fossil [was: List of companies that use Plan 9.]

[Noam Preil]

Responding off list shortly :)

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T6b867aa3be7bf660-M3fe517cc779e245e44a024b1
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Balancing Progress and Accessibility in the Plan 9 Community. (Was: [9fans] Interoperating between 9legacy and 9front)

[Noam Preil]

There's a clear pattern, though. The document is blatantly AI-generated,
and I believe that the author acknowledged it as such ("the model was
confirmed as trained on 9front sources"); even if it wasn't, the logical
mistakes it makes are of a type humans don't generally make.

The author has many posts that _all_ feature A.I. art. The arguments
they make have no connection to the premises, although in fairness
that's a hallmark of bad human writing too. The sources cited have no
connection to the arguments being made.

It is not unreasonable to assume that someone who is clearly relying so
heavily on LLMs might be doing so on the mailing list when so many of
their posts clearly resemble LLM output.

Hell, even if they're not using an LLM, if someone is writing with the
_quality_ of an LLM, they're not worth engaging with. "This horrible
spam-looking content was actually written by a person!" is not a great
defense.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Te051f230f2656bbb-M50b020de4168865f586f0fdb
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] List of companies that use Plan 9.

[Noam Preil]

I have found multiple deadlocks over the last few years, and only
bothered fixing one of them. That patch is in 9legacy, as well, now.

During my IWP9 talk in which, among other things, I explained why I
intend to replace fossil despite curently using it, I demonstrated one
of the problems with fossil by (attempting to) install Go, which crashes
the file system _every single time_.

I personally fixed over a dozen issues in Venti, and continue to find
more. To the best of my knowledge, those patches are currently _only_
present in 9front. I was not aware 9legacy existed when I did that
work.

I've lost some data to fossil crashes, since if it e.g. loses power
during a sync, it will end up _erasing_ a file rather than preserving
the old version, which has _broken boot_ if the system crashed while I
was editing lib/profile.

I have a branch of 9front with fossil restored. I have discussed what
would be needed to add fossil upstream again, and there was acceptance
to the possibility if the issues get worked out _first_.

The hostility on this list is pervasive and comes from all sides. If
your goal is positive engagement, this is not the place to do it.



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tad3dc0c93039a7d2-Me721f99709008ac03b77ea0b
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] List of companies that use Plan 9.

[Noam Preil]

The only company listed there is yours, no?

That's not a good look.



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tad3dc0c93039a7d2-M75d1829e0bc8637cd5edecf5
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Balancing Progress and Accessibility in the Plan 9 Community. (Was: [9fans] Interoperating between 9legacy and 9front)

[Noam Preil]

It is not, but in fairness, some people really do write the same way
that the LLMs do, so it's not impossible for a real person to appear to
be an LLM.

I'm leaning towards LLM-generated, anyways. It's far too similar in
structure / syntax to other AI spam I've had to deal with recently.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Te051f230f2656bbb-M6b07d6fc6ad01c998e34b8ed
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Interoperating between 9legacy and 9front

[Noam Preil]

> The absence of Fossil from 9front was the one I found most difficult to
> overcome, but at least in theory only the equivalent of "fossil/conf" (an
> rc script I eventually shoehorned from plan9port) is essential. I can see
> how it would be inconvenient to need to support software that is
> significantly complex, especially when it must also be able to be embedded
> in the kernel.

Actually, adding fossil back in to 9front is extremely simple; I have a
branch at https://git.sr.ht/~pixelherodev/plan9 which has fossil
integrated.

The talk I gave at IWP9 was running from fossil on my 9front branch.

If my changes are too extensive compared to 9front (it's a personal
branch, so I wouldn't blame you for thinking that), I'm happy to even
create a branch that's just 9front+fossil. It's really not hard.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tde2ca2adda383a3a-Mf4487e7706a5d7b4363ebe2e
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] Balancing Progress and Accessibility in the Plan 9 Community. (Was: [9fans] Interoperating between 9legacy and 9front)

[Noam Preil]

A document that had no research put into it is not a first draft, it's
at best spam.

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Te051f230f2656bbb-M636e86d3c30c32c5c186e579
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[noam]

Quoth Edouard Klein :
> I, for one, will attend, barring any incident.

Same!

> I will send my submission in a frenzy panic minutes before the deadline,
> as one usually does.

Hey, same!! :D


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-M3ae2c416df89b4067059249b
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[noam]

Quoth Don Bailey :
> Cool rant.

It's incredible to me that you can put in so much effort to make sure that
the strength of your convictions is clear, while also putting in so little
effort that nobody wants to be around you long enough to find out what they
actually are.


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-M9ed40c01efb5cf9a09f74622
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[noam]

Quoth Don Bailey :
> I'm not sure what all this was, so I didn't read most of it.

Are you okay?


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-M74c1edb8077ff8ea2c855ac1
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[Noam Preil]

I wasn't talking to you :P

That was a response to the post that looked AI-generated. I sincerely
apologize if I accidentally responded to your post instead, but I don't
think I did

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-Md8d4493d987b88a15909254c
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[Noam Preil]

That seems like a lot of work, though; surely, we can all just share a
rootfs over 9p!!


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-M599b7f6fddada07bc414d828
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community. [Was:Re: [9fans] Supported Notebooks]

[Noam Preil]

It might help to believe that if you wrote actual ideas about what
should be done.

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-M42002f4830a2e854bd1548f5
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

[9fans] Re: Charting the Future: Envisioning Plan 9 Release 5 for the 9fans Community.

[noam]

Quoth vic.thac...@fastmail.fm:
> [redacted for sanity]

This is AI-generated spam, right? :P

- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T42f11e0265bcfa18-Mfbfde21dd14cb18b95481013
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] 9front: fossil broken in Humanbiologics rel.

[Noam Preil]

Looks like it's hanging waiting for a response from venti; are you able to 
connect to venti directly? Did your venti or fossil config change recently? I'm 
not aware of recent changes to upstream venti that could cause this, but it's 
certainly possible, and venti and fossil are both fragile enough that i can 
imagine recent libc changes breaking them.

This also explains why Ori and Cinap experienced no issues, since their tests 
did not involve venti connections.

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tac8d983292c826c1-Mf3cc3500a17468c2f39f3793
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] 9front: fossil broken in Humanbiologics rel.

[Noam Preil]

I run venti+fossil off of 9front git without issues (well, without issues 
caused by 9front ;).

Can you give a full stack trace of all threads? Happy to take a look :)

- Noam Preil

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tac8d983292c826c1-Mdb07f7bd2a0bf83b4af51a8e
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] 9Legacy status

[noam]

Quoth Don Bailey :
> Hi All,
> 
> Just curious if 9Legacy is considered the right place to submit modern
> patches? Is that kind of the 'go to' repository for modern
> patches/improvements to the 'mainline' Bell Labs distro?

https://9front.org



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T53b730e0213e9325-M4e56567220bd8d2455fa5b53
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] [PATCH] fossil: fix a deadlock in the caching logic

[noam]

Quoth Charles Forsyth :
> fussing about certain things for hard drives that probably don't matter for
> SSD let alone nvme
> certain things

I am once again asking you to be more specific, please :)

I have Plans for improving venti for myself, it'd be great to actually
have a specific list of issues that others have noticed!

Thanks,


- Noam Preil


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T354fe702e1e9d5e9-M9ba84777a354c147540a1e76
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] [PATCH] fossil: fix a deadlock in the caching logic

[noam]

Quoth Anthony Martin :
> Noam, can you reproduce your problem with the above patch?

Just from reading the patch: yes, with ~95% confidence.
The logic it's fixing is completely unrelated.

So: thanks! Now I can import that patch and have *another* deadlock fixed!

Are there any other patches I should know about?



--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T354fe702e1e9d5e9-M888aed33d63b46d671dc5ba0
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] [PATCH] fossil: fix a deadlock in the caching logic

[noam]

Quoth Charles Forsyth :
> Fossil works hard to do certain things that now we probably wouldn't bother
> to do.

Such as?


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T354fe702e1e9d5e9-M8a784194389d102772df0a9e
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] [PATCH] fossil: fix a deadlock in the caching logic

[noam]

Quoth Steve Simon :
> was this hard to reproduce?

I've seen it sporadically over the last year, and - when looking for it -
was able to trigger it deliberately yesterday with minimal effort the
first time I looked for it.

...of course, when I was later trying to reproduce it a third time, I
wasn't able to trigger it even on the fs running *without* the patch, so.
~50% reproducibility rate so far when I'm actively trying to hit it.

I'm fairly sure the root cause is a race condition between some of the
periodic threads - this is only triggered when we try to flush a clean
block, which isn't a common occurrence - but I wouldn't have put in so
much effort to fix this if it wasn't something I semiregularly ran into.

for(f in `{walk /sys})
chmod +w $f

I think this, combined with the periodic flush routines and tight timing,
consistently reproduces it. There's probably a more general way to do so,
but without diving even deeper and seeing how we end up trying to flush
clean blocks, it's hard to say for sure.

fossil is *usually* pretty stable for me these days. My thinkpad often
has an uptime of weeks, and usually resets because I'm hacking on the
system and need to reboot to test it, not because of fossil.

I *have* seen bugs depressingly often, though. Once every month or so,
pretty consistently.

e.g. building any version of Go newer than 1.7 on my thinkpad crashes
fossil with ~50% consistency. The rest of the time, it alternates
between sporadic failures due to bugs in the Go compiler, and actually
working.

Similarly, using kvik's clone tool to move large volumes of data has
been a reliable way to crash fossil for me in the past (100% reliability
- there was an invocation that would cause the system to die literally
every time, but I don't remember exactly which dataset it was, or
what level of parallelism was required).


--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T354fe702e1e9d5e9-M4de9844d7316ae1c346391d7
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

[9fans] [PATCH] fossil: fix a deadlock in the caching logic

[noam]

I've sporadically encountered a deadlock in fossil. Naturally, when your root 
file system crashes, it can be hard to debug. My solution: stop having a root 
file system. Was able to attach acid using mycroft's tooling from ANTS, and get 
a clean stack trace (https://pixelhero.dev/notebook/fossil/stacks/2023-04-03.1).

After a few hours yesterday 
(https://pixelhero.dev/notebook/fossil/2023-04-03.html 
), I eventually tracked down the 
deadlock. When blockWrite is told to flush a clean block to disk - i.e. one 
which is already flushed - it removes the block from the cache's free list, 
locks the block, detects that it's clean, and then... drops the reference. 
While keeping the block locked. And in the cache.

This leak of the lock, of course, means that the *next* access to the block - 
which is still in the cache! - hangs indefinitely. This is seen exactly in the 
stack trace:

_cacheLocal grabs the block from the cache, tries to lock it, and hangs 
indefinitely. Worse, it does so under a call to fileWalk, which holds a 
different lock, so the effect spreads out and makes even more of the file 
system inaccessible as well (the fileMetaFlush proc hangs waiting on this file 
lock).

This patch just ensures we call blockPut on the BioClean path as well, thus 
unlocking the block and readding it to the cache's free lists.

The patch is on my branch - 
https://git.sr.ht/~pixelherodev/plan9/commit/1bf8bd4f44e058261da7e89d87527b12073c9e0f
 - but I figured I should probably post it here as well.

If anyone has any other patches that weren't in the 9legacy download as of 
~2018, please let me know! :)

---
sys/src/cmd/fossil/cache.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/sys/src/cmd/fossil/cache.c b/sys/src/cmd/fossil/cache.c
index f473d211e..2fec44949 100644
--- a/sys/src/cmd/fossil/cache.c
+++ b/sys/src/cmd/fossil/cache.c
@@ -1203,8 +1203,10 @@ blockWrite(Block *b, int waitlock)
fprint(2, "%s: %d:%x:%d iostate is %d in blockWrite\n",
argv0, bb->part, bb->addr, bb->l.type, bb->iostate);
/* probably BioWriting if it happens? */
-   if(bb->iostate == BioClean)
+   if(bb->iostate == BioClean){
+   blockPut(bb);
goto ignblock;
+   }
}

blockPut(bb);
--

--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T354fe702e1e9d5e9-Ma4bc3ee3ea80defbb239f382
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Re: [9fans] p9f mention of 9front

[noam]

I'm working on rebasing ANTS onto 9front-master, with the intent of taking up 
maintainance.
--
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/T523d6e906a17a7cc-Me752f3cc22bb8ccd5a0bc381
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription