From jondrews at fastmail.com  Fri Dec  3 23:51:09 2021
From: jondrews at fastmail.com (Jonathan Drews)
Date: Fri, 3 Dec 2021 21:51:09 -0700
Subject: [Semibug] Use mount_mfs to speed up programs
In-Reply-To: <935b6b21-f8d0-4c0c-7661-7ec489222586@holland-consulting.net>
References: <YaXoGk6NA/DABIht@leo.easyfashionclothing.com>
 <935b6b21-f8d0-4c0c-7661-7ec489222586@holland-consulting.net>
Message-ID: <YarzvWPMci9rLyuT@leo.easyfashionclothing.com>

On Tue, Nov 30, 2021 at 08:15:21AM -0500, Nick Holland wrote:
> On 11/30/21 4:00 AM, Jonathan Drews wrote:
> > Hi Folks:
> > 
> >   I used Michael's book "Absolute OpenBSD" to setup MFS (Memory File
> > System). It is also in man 8 mount_mfs. I did:
> > 
> > # mount_mfs -s 2048m /dev/sd0b /mnt
> 
> Quoting D. J. Bernstein: "Benchmark, don't speculate"
> 
> Having an MFS be 100x faster than your hard drive only matters if
> your hard drive is a significant bottleneck.
> 
> If your program does little from the disk, you are more likely to
> spend more time setting up the loading of the MFS than you will gain
> from program execution.  And a LOT more time recreating data /when/
> you forget to put the data back on real disk when you are done with

Nick and Michael:

 Thanks for the advice on this. The MFS was interesting to play with but
I will not use it. 

--
Kind regards,
Jonathan


From mmattox at support.tools  Sun Dec  5 15:16:22 2021
From: mmattox at support.tools (Matthew Mattox)
Date: Sun, 5 Dec 2021 14:16:22 -0600
Subject: [Semibug] Semibug Digest, Vol 64, Issue 1
In-Reply-To: <mailman.3.1638723602.98599.semibug@lists.nycbug.org>
References: <mailman.3.1638723602.98599.semibug@lists.nycbug.org>
Message-ID: <65E25304-A8DB-4D67-9610-A1A1AE9B1939@support.tools>

Question, is mount_mfs different than a RAM disk (mount -t tmpfs -o size=1024m myramdisk /tmp/ramdisk) ?

Sent from my iPhone

> On Dec 5, 2021, at 11:00, semibug-request at lists.nycbug.org wrote:
> 
> ?Send Semibug mailing list submissions to
>    semibug at lists.nycbug.org
> 
> To subscribe or unsubscribe via the World Wide Web, visit
>    http://lists.nycbug.org:8080/mailman/listinfo/semibug
> or, via email, send a message with subject or body 'help' to
>    semibug-request at lists.nycbug.org
> 
> You can reach the person managing the list at
>    semibug-owner at lists.nycbug.org
> 
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Semibug digest..."
> 
> 
> Today's Topics:
> 
>   1. Re: Use mount_mfs to speed up programs (Jonathan Drews)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Fri, 3 Dec 2021 21:51:09 -0700
> From: Jonathan Drews <jondrews at fastmail.com>
> To: semibug at lists.nycbug.org
> Subject: Re: [Semibug] Use mount_mfs to speed up programs
> Message-ID: <YarzvWPMci9rLyuT at leo.easyfashionclothing.com>
> Content-Type: text/plain; charset=us-ascii
> 
>> On Tue, Nov 30, 2021 at 08:15:21AM -0500, Nick Holland wrote:
>>> On 11/30/21 4:00 AM, Jonathan Drews wrote:
>>> Hi Folks:
>>> 
>>>  I used Michael's book "Absolute OpenBSD" to setup MFS (Memory File
>>> System). It is also in man 8 mount_mfs. I did:
>>> 
>>> # mount_mfs -s 2048m /dev/sd0b /mnt
>> 
>> Quoting D. J. Bernstein: "Benchmark, don't speculate"
>> 
>> Having an MFS be 100x faster than your hard drive only matters if
>> your hard drive is a significant bottleneck.
>> 
>> If your program does little from the disk, you are more likely to
>> spend more time setting up the loading of the MFS than you will gain
>> from program execution.  And a LOT more time recreating data /when/
>> you forget to put the data back on real disk when you are done with
> 
> Nick and Michael:
> 
> Thanks for the advice on this. The MFS was interesting to play with but
> I will not use it. 
> 
> --
> Kind regards,
> Jonathan
> 
> 
> 
> ------------------------------
> 
> Subject: Digest Footer
> 
> _______________________________________________
> Semibug mailing list
> Semibug at lists.nycbug.org
> http://lists.nycbug.org:8080/mailman/listinfo/semibug
> 
> 
> ------------------------------
> 
> End of Semibug Digest, Vol 64, Issue 1
> **************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.nycbug.org:8443/pipermail/semibug/attachments/20211205/5d7f72ff/attachment.htm>

From nick at holland-consulting.net  Mon Dec  6 08:17:44 2021
From: nick at holland-consulting.net (Nick Holland)
Date: Mon, 6 Dec 2021 08:17:44 -0500
Subject: [Semibug] Semibug Digest, Vol 64, Issue 1
In-Reply-To: <65E25304-A8DB-4D67-9610-A1A1AE9B1939@support.tools>
References: <mailman.3.1638723602.98599.semibug@lists.nycbug.org>
 <65E25304-A8DB-4D67-9610-A1A1AE9B1939@support.tools>
Message-ID: <a537dbbb-7241-ca24-5149-23756466d160@holland-consulting.net>

yes...
As I recall, the mfs code was considered somewhat old and crufty,
and tmpfs was hoped to be a replacement.  However, there were
bugs in the new code and no one was actively working on them, so
at  one point, tmpfs was disabled.  It seems while mount_tmpfs is
still built, it is disabled in the kernel, the idea is if someone
wants to step up and make it work better, it is there.

Personally, I had used it a fair amount, never saw issues with it,
but mfs works fine for me, too.

Nick.

On 12/5/21 3:16 PM, Matthew Mattox wrote:
> Question, is mount_mfs different than a RAM disk (mount -t tmpfs -o size=1024m myramdisk /tmp/ramdisk) ?
> 
> Sent from my iPhone
> 
>> On Dec 5, 2021, at 11:00, semibug-request at lists.nycbug.org wrote:
>>
>> ?Send Semibug mailing list submissions to
>> ? ?semibug at lists.nycbug.org
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>> ? ?http://lists.nycbug.org:8080/mailman/listinfo/semibug
>> or, via email, send a message with subject or body 'help' to
>> ? ?semibug-request at lists.nycbug.org
>>
>> You can reach the person managing the list at
>> ? ?semibug-owner at lists.nycbug.org
>>
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of Semibug digest..."
>>
>>
>> Today's Topics:
>>
>> ??1. Re: Use mount_mfs to speed up programs (Jonathan Drews)
>>
>>
>> ----------------------------------------------------------------------
>>
>> Message: 1
>> Date: Fri, 3 Dec 2021 21:51:09 -0700
>> From: Jonathan Drews <jondrews at fastmail.com>
>> To: semibug at lists.nycbug.org
>> Subject: Re: [Semibug] Use mount_mfs to speed up programs
>> Message-ID: <YarzvWPMci9rLyuT at leo.easyfashionclothing.com>
>> Content-Type: text/plain; charset=us-ascii
>>
>> On Tue, Nov 30, 2021 at 08:15:21AM -0500, Nick Holland wrote:
>>> On 11/30/21 4:00 AM, Jonathan Drews wrote:
>>>> Hi Folks:
>>>>
>>>> ?I used Michael's book "Absolute OpenBSD" to setup MFS (Memory File
>>>> System). It is also in man 8 mount_mfs. I did:
>>>>
>>>> # mount_mfs -s 2048m /dev/sd0b /mnt
>>>
>>> Quoting D. J. Bernstein: "Benchmark, don't speculate"
>>>
>>> Having an MFS be 100x faster than your hard drive only matters if
>>> your hard drive is a significant bottleneck.
>>>
>>> If your program does little from the disk, you are more likely to
>>> spend more time setting up the loading of the MFS than you will gain
>>> from program execution. ?And a LOT more time recreating data /when/
>>> you forget to put the data back on real disk when you are done with
>>
>> Nick and Michael:
>>
>> Thanks for the advice on this. The MFS was interesting to play with but
>> I will not use it.
>>
>> --
>> Kind regards,
>> Jonathan
>>
>>
>>
>> ------------------------------
>>
>> Subject: Digest Footer
>>
>> _______________________________________________
>> Semibug mailing list
>> Semibug at lists.nycbug.org
>> http://lists.nycbug.org:8080/mailman/listinfo/semibug
>>
>>
>> ------------------------------
>>
>> End of Semibug Digest, Vol 64, Issue 1
>> **************************************
> 
> _______________________________________________
> Semibug mailing list
> Semibug at lists.nycbug.org
> https://lists.nycbug.org:8443/mailman/listinfo/semibug
> 


From jondrews at fastmail.com  Mon Dec 20 10:34:22 2021
From: jondrews at fastmail.com (Jonathan Drews)
Date: Mon, 20 Dec 2021 08:34:22 -0700
Subject: [Semibug] Meeting for 21 Dec and Future Talks
Message-ID: <YcCifilcj5dlp3ki@leo.easyfashionclothing.com>

Hi Folks:

 Tomorrow, 21 December, is our monthly meetup. Here is the link for
Jitsi:

https://meet.jit.si/Semi-BUGDecemberMeeting

The meeting begins at 1900 Detroit, MI time. 

For the new year I have the folowing presentations lined up:

January 2022: Alan Jude will give a talk on The ZFS file system.

February 2022: Susan Hurst, an experienced DB administrator will give
a talk on her new book, "The Left Side of Monday". It is about the
problems in translating customer (and boses) requests into robust
databases.

March 2022: Ed Howland, a programmer with 23 years experience, will
give a presentation on Backus-Naur Form and how it can help you in
your programming.
An overview of Backus Naur
https://isaaccomputerscience.org/concepts/dsa_toc_bnf?examBoard=all&stage=all

Let me know what you think of these topics.  

Kind regards,

Jonathan


From markmoellering at psyberation.com  Thu Dec 30 15:10:42 2021
From: markmoellering at psyberation.com (Mark Moellering)
Date: Thu, 30 Dec 2021 15:10:42 -0500
Subject: [Semibug] Question on max VDEVs in Zpools
Message-ID: <090c79cb-c156-340b-a21a-afa3fb4ae7e8@psyberation.com>

Question for the group / Michael

Is there a? limit to the number of VDEVs I can put in a zpool? If I want 
a huge zpool, can I put 100+ VDEVs into a zpool?? Trying to figure out 
practical limits.

Thanks


-- mark


From njt at ayvali.org  Fri Dec 31 04:48:26 2021
From: njt at ayvali.org (N.J. Thomas)
Date: Fri, 31 Dec 2021 01:48:26 -0800
Subject: [Semibug] Question on max VDEVs in Zpools
In-Reply-To: <090c79cb-c156-340b-a21a-afa3fb4ae7e8@psyberation.com>
References: <090c79cb-c156-340b-a21a-afa3fb4ae7e8@psyberation.com>
Message-ID: <Yc7R6jSE8lses0Yo@ayvali.org>

* Mark Moellering <markmoellering at psyberation.com> [2021-12-30 15:10:42-0500]:
> Is there a?? limit to the number of VDEVs I can put in a zpool? If I
> want a huge zpool, can I put 100+ VDEVs into a zpool??? Trying to
> figure out practical limits.

Interesting question. If I am reading them correctly, a number of
documents online (including the Wikipedia ZFS entry:
https://en.wikipedia.org/wiki/ZFS#Capacity) claim this to be 2^64
("number of devices in any zpool"), but I couldn't find hard evidence of
this in the OpenZFS source code as I quickly perused through it. (Keep
in mind I was just grepping casually though.)

hth,
Thomas