mysqlfs-general — General discussion list for MySQLfs users and developers.

Re: [Mysqlfs-general] About creating lowlevel version of myslqfs.

[Stef B. <st...@bo...>]

  On 11/10/2010 07:49 PM, Allan Clark wrote:
> On Wed, Nov 10, 2010 at 07:48, Stef Bon<st...@bo...>  wrote:
>>   Hello,
>>
>> has somebody tried to create a lowlevel version of mysqlfs?
> Hi Stef;
>
> What do you mean by a low-level version?  It's currently a FUSEfs, how
> much lower do you mean?

Well, fuse offers two interfaces you can make use of to make your fs.

A highlevel, where the path is an argument, and a lowlevel one, where 
the inode is an argument.

I've started to program the highlevel interface, which is easier, and 
switched to the lowlevel,
cause I want to implement aio with my fs, and that is easier with the 
low level interface.

See:

http://fuse.sourceforge.net/doxygen/structfuse__lowlevel__ops.html

So there are different level of "lowness"...

Stef

Re: [Mysqlfs-general] About creating lowlevel version of myslqfs.

[Allan C. <al...@ch...>]

On Wed, Nov 10, 2010 at 07:48, Stef Bon <st...@bo...> wrote:
>  Hello,
>
> has somebody tried to create a lowlevel version of mysqlfs?

Hi Stef;

What do you mean by a low-level version?  It's currently a FUSEfs, how
much lower do you mean?

Allan
-- 
al...@ch...  "金鱼" http://linkedin.com/in/goldfish
please, no proprietary attachments (http://tinyurl.com/cbgq)

[Mysqlfs-general] About creating lowlevel version of myslqfs.

[Stef B. <st...@bo...>]

  Hello,

has somebody tried to create a lowlevel version of mysqlfs?


Stef

Re: [Mysqlfs-general] ./configure pthreads

[Michal L. <mi...@lo...>]

Blakey wrote:
> Hi Everyone,
> 
> I'm just trying to compile mysqlfs, but have a minor issue, ./configure 
> seems to bomb out once it gets to the pthreads testing section.

IIRC it's got something to do with MySQL or FUSE libraries. Although it
says pthreads lib is missing it actually means libmysql-dev or
libfuse-dev (or how are these called on Ubuntu) are missing. Make sure
you've got them installed and re-run configure. It should work.

Michal

[Mysqlfs-general] ./configure pthreads

[Blakey <bl...@bo...>]

Hi Everyone,

I'm just trying to compile mysqlfs, but have a minor issue, ./configure 
seems to bomb out once it gets to the pthreads testing section.

I've had a search around and it would appear that the pthreads library 
is installed.

Does anyone have any ideas / suggestions?

I'm running x86 xubuntu 8.04 if thats of any use...

Thanks,
Craig

Re: [Mysqlfs-general] Seperate the servers to allow different insert/update hosts.

[Smart W. G. F. W. <f.w...@sm...>]

Hi,

Shane Spencer schrieb:
> So, I set up mysql master to slave replication as a test.  It works
> really stupidly well with mysqlfs for distributing an instant backup
> offsite somewhere.  I'm hoping to use mysqlfs for some awesome,
> pointless, evil..
> 
> Any chance you could aid my effort by allowing two DB connections if
> requested?  One for insert/update and one for all selects?  This would
> allow changes to go upstream (as well as local w/ good locking).  I
> know this probably makes your brain twist a bit, but c'mon.  It'll be
> unsafe, unprotected byte-sex.  You know you wanna do it.


You can just use mysql-ndbcluster for that purpose and mount mysqlfs 
readonly on the second client.


-- 

Mit freundlichen Grüßen,

Florian Wiessner

Smart Weblications GmbH
Martinsberger Str. 1
D-95119 Naila

fon.: +49 700 762 789 32 - 0,12 EUR/Min*
fax.: +49 700 762 789 32 - 0,12 EUR/Min*
24/7: +49 900 311 886 00 - 1,99 EUR/Min*
http://www.smart-weblications.de

--
Sitz der Gesellschaft: Naila
Geschäftsführer: Florian Wiessner
HRB-Nr.: HRB 3840 Amtsgericht Hof
*aus dem dt. Festnetz, ggf. abweichende Preise aus dem Mobilfunknetz

Re: [Mysqlfs-general] Seperate the servers to allow different insert/update hosts.

[Andrew R. <ros...@gm...>]

On 14/04/2008, Shane Spencer <sh...@bo...> wrote:
> So, I set up mysql master to slave replication as a test.  It works
>  really stupidly well with mysqlfs for distributing an instant backup
>  offsite somewhere.  I'm hoping to use mysqlfs for some awesome,
>  pointless, evil..
>
>  Any chance you could aid my effort by allowing two DB connections if
>  requested?  One for insert/update and one for all selects?  This would
>  allow changes to go upstream (as well as local w/ good locking).  I
>  know this probably makes your brain twist a bit, but c'mon.  It'll be
>  unsafe, unprotected byte-sex.  You know you wanna do it.


Hi Shane

 Why not use mysql proxy (http://forge.mysql.com/wiki/MySQL_Proxy)
first just to be sure what your doing actually requires "unprotected
byte-sex" ;)

-
Andrew Rose
http://andrew-rose.blogspot.com

[Mysqlfs-general] Seperate the servers to allow different insert/update hosts.

[Shane S. <sh...@bo...>]

So, I set up mysql master to slave replication as a test.  It works
really stupidly well with mysqlfs for distributing an instant backup
offsite somewhere.  I'm hoping to use mysqlfs for some awesome,
pointless, evil..

Any chance you could aid my effort by allowing two DB connections if
requested?  One for insert/update and one for all selects?  This would
allow changes to go upstream (as well as local w/ good locking).  I
know this probably makes your brain twist a bit, but c'mon.  It'll be
unsafe, unprotected byte-sex.  You know you wanna do it.

- Shane

Re: [Mysqlfs-general] query_inode_full: how does it work?

[Smart W. G. F. W. <f.w...@sm...>]

Hi,

Stef Bon schrieb:

>> Yes, the query grows bigger, and it is executed every time a lstat() or 
>> lstat64() syscall happens - which makes performance really bad on 
>> ndbcluster with distributed clients on different locations. We should 
>> cache it somewhere (either in mysqlfs_getattr, or as cache somewhere in 
>> query_inode_full.
>>
>> the cache must be cleared for the inode, if write/rename or chown/chmod 
>> is executed so that query_inode_full always returns proper values.
>>
>> i thought about that inode-cache - it shouldn't be more than an array 
>> containing values which has to be checked before the query is run.
>>   
> Yes, a cache should will things up. The construction now does a lot of 
> lookups and never uses
> data it has found before. For example, if you want to read a file 
> /a/b/c/d/testfile, it will check you've permissions to
> that file, and walk the path to it:
> check /
[check...]
> check /a/b/c/d/testfile
> 
> to determine the inode of say /a/b/c, it does not use the information it 
> just found before.
> 
> So a cache is a smart thing to do.
> There are other fuse fs's using a cache(sshfs, fusesmb). My idea is to 
> look first there how it is solved there.
> My first impression is that fuse offers possibilities to use a cache, 
> but I'm not sure.
> 
>> as for the locking-problem, what about creating a table metadata to 
>> store information on which inode is currently locked? metadata could 
>> also be stored with ndbcluster-engine, and thus we would easily be able 
>> to implement file locking.
>>
>>   
> Why not create an extra field in the inodes table, with a information 
> about the locking?
> An extra table is making things extra complicated, I think.
> 

ok, i'll look at other fuse-fs's how cache is implemented there.

one extra field in the inode table is sufficient if mysqlfs is not run 
on a cluster, but i don't know if it would be smarter to store this kind 
of metainfo in a seperate table "metadata" - e.g. which mysql-host has 
locked which inode etc.

but i think we should first implement the cache, then we can solve the 
locking problem...



-- 

Mit freundlichen Grüßen,

Florian Wiessner

Smart Weblications GmbH
Martinsberger Str. 1
D-95119 Naila

fon.: +49 700 762 789 32 - 0,12 EUR/Min*
fax.: +49 700 762 789 32 - 0,12 EUR/Min*
24/7: +49 900 311 886 00 - 1,99 EUR/Min*
http://www.smart-weblications.de

--
Sitz der Gesellschaft: Naila
Geschäftsführer: Florian Wiessner
HRB-Nr.: HRB 3840 Amtsgericht Hof
*aus dem dt. Festnetz, ggf. abweichende Preise aus dem Mobilfunknetz

Re: [Mysqlfs-general] query_inode_full: how does it work?

[Stef B. <st...@bo...>]

Smart Weblications GmbH, Florian Wiessner wrote:
> Hi,
>
> Stef Bon schrieb:
>
>   
>> This query does grow bigger as the path is longer (deeper).
>>
>>     
> Yes, the query grows bigger, and it is executed every time a lstat() or 
> lstat64() syscall happens - which makes performance really bad on 
> ndbcluster with distributed clients on different locations. We should 
> cache it somewhere (either in mysqlfs_getattr, or as cache somewhere in 
> query_inode_full.
>
> the cache must be cleared for the inode, if write/rename or chown/chmod 
> is executed so that query_inode_full always returns proper values.
>
> i thought about that inode-cache - it shouldn't be more than an array 
> containing values which has to be checked before the query is run.
>   
Yes, a cache should will things up. The construction now does a lot of 
lookups and never uses
data it has found before. For example, if you want to read a file 
/a/b/c/d/testfile, it will check you've permissions to
that file, and walk the path to it:
check /
check /a
check /a/b
check /a/b/c
check /a/b/c/d
check /a/b/c/d/testfile

to determine the inode of say /a/b/c, it does not use the information it 
just found before.

So a cache is a smart thing to do.
There are other fuse fs's using a cache(sshfs, fusesmb). My idea is to 
look first there how it is solved there.
My first impression is that fuse offers possibilities to use a cache, 
but I'm not sure.


> as for the locking-problem, what about creating a table metadata to 
> store information on which inode is currently locked? metadata could 
> also be stored with ndbcluster-engine, and thus we would easily be able 
> to implement file locking.
>
>   
Why not create an extra field in the inodes table, with a information 
about the locking?
An extra table is making things extra complicated, I think.

Stef Bon

[Mysqlfs-general] mysqlfs tests and timings

[Smart W. G. F. W. <f.w...@sm...>]

hi all,


i have done the following setup:

server1: amd athlon 64 x2, 2gb ram, 2x250gb hdd software raid1
server2: intel p4 1.8ghz, 256mb ram, 1x20gb ide disk

the servers are connected via 100mbit, but with async routing over 7 
hops with about 10ms latency.

i have mounted a mysqlfs with ndb-cluster and put my website on it 
(www2.smart-weblications.de is now dns-round-robin with mysqlfs 
underneath), to benchmark the performance:

http://www2.smart-weblications.de/mysqlfs-strace.txt

it takes about 6-8 seconds to render the page - most of the time the 
system waits on lstat64() (which is query_inode_full() in our case...)

i also copied files from and to the mysqlfs mountpoint:

time cp mysqlfs-strace.txt /tmp/foo

real    0m0.394s
user    0m0.002s
sys     0m0.001s

time cp /tmp/strace4 mysqlfs-strace.txt

real    0m5.677s
user    0m0.001s
sys     0m0.002s

which means that reading from the fs is quite fast (approx. 700kb/sek), 
writing is awful slow (approx. 10kb/sek)...

http://www2.smart-weblications.de/mysqlfs-strace-write.txt

so somehow we should cache write()

-- 

Mit freundlichen Grüßen,

Florian Wiessner

Smart Weblications GmbH
Martinsberger Str. 1
D-95119 Naila

fon.: +49 700 762 789 32 - 0,12 EUR/Min*
fax.: +49 700 762 789 32 - 0,12 EUR/Min*
24/7: +49 900 311 886 00 - 1,99 EUR/Min*
http://www.smart-weblications.de

--
Sitz der Gesellschaft: Naila
Geschäftsführer: Florian Wiessner
HRB-Nr.: HRB 3840 Amtsgericht Hof
*aus dem dt. Festnetz, ggf. abweichende Preise aus dem Mobilfunknetz

Re: [Mysqlfs-general] query_inode_full: how does it work?

[Smart W. G. F. W. <f.w...@sm...>]

Hi,

Stef Bon schrieb:

>>
>> now I can check the flow of what happens (by activating debugging). It's 
>> much more clear now.
>>
>> But when I look at query_inode_full, I do not understand what happens 
>> there at all.
>> The purpose of this function is to translate a path into a inode, right? 
>> But it does much more things.
>> It "returns" a table nlinks, the inode, the parent. (and more?)
>>
>> Can someone explain why this is? It runs complicated queries....
>>   
> 
> Ok,
> 
> now I've read it a couple of times, I understand it more or less.
> 
> A (long) query is constructed to determine the inode, with the help of a 
> LEFT JOIN.
> 
> This query does grow bigger as the path is longer (deeper).
> 
> I'm looking into the mysql syntax now. Any hints about it welcome.
> 

Yes, the query grows bigger, and it is executed every time a lstat() or 
lstat64() syscall happens - which makes performance really bad on 
ndbcluster with distributed clients on different locations. We should 
cache it somewhere (either in mysqlfs_getattr, or as cache somewhere in 
query_inode_full.

the cache must be cleared for the inode, if write/rename or chown/chmod 
is executed so that query_inode_full always returns proper values.

i thought about that inode-cache - it shouldn't be more than an array 
containing values which has to be checked before the query is run.

as for the locking-problem, what about creating a table metadata to 
store information on which inode is currently locked? metadata could 
also be stored with ndbcluster-engine, and thus we would easily be able 
to implement file locking.

-- 

Mit freundlichen Grüßen,

Florian Wiessner

Smart Weblications GmbH
Martinsberger Str. 1
D-95119 Naila

fon.: +49 700 762 789 32 - 0,12 EUR/Min*
fax.: +49 700 762 789 32 - 0,12 EUR/Min*
24/7: +49 900 311 886 00 - 1,99 EUR/Min*
http://www.smart-weblications.de

--
Sitz der Gesellschaft: Naila
Geschäftsführer: Florian Wiessner
HRB-Nr.: HRB 3840 Amtsgericht Hof
*aus dem dt. Festnetz, ggf. abweichende Preise aus dem Mobilfunknetz

Re: [Mysqlfs-general] Maintaining previous versions.

[Stef B. <st...@bo...>]

Michal Ludvig wrote:
> I think what you really want is something like "rdiff-backup" or 
> "duplicity". Both tools let you maintain backups with history, remove 
> old revisions, etc. You can indeed store the archives in mysqlfs if you 
> wish.
>
> Have a look at:
> - http://www.nongnu.org/duplicity/
> - http://rdiff-backup.stanford.edu/
> - http://www.debian-administration.org/articles/209 (not really 
> debian-specific, don't worry)
>   
Thanks for the links. I'm using rdiff-backup already on some machines now.

The reason for trying to do this with mysqlfs is that it has special 
potential
on maintaining versions, cause it's fuse and it's in a database. Then 
you do not need a
complicated script as rdiff-backup, but backups are made on the fly.

I'll check the links anyway.

Stef Bon

Re: [Mysqlfs-general] Maintaining previous versions.

[Michal L. <mi...@lo...>]

Stef Bon wrote:

> My idea is to only do a copy or a sync of a directory of my choice, for 
> example my homedirectory, once a day (4.30 at night for example), to the 
> mounted mysqlfs.
> This way not every change will get his way to the database.
> 
> Of course there should be a policy on how long a version will stay in 
> the database, and there should be some utilities to find previous 
> versions. But that's not the most difficult step I think.

I think what you really want is something like "rdiff-backup" or 
"duplicity". Both tools let you maintain backups with history, remove 
old revisions, etc. You can indeed store the archives in mysqlfs if you 
wish.

Have a look at:
- http://www.nongnu.org/duplicity/
- http://rdiff-backup.stanford.edu/
- http://www.debian-administration.org/articles/209 (not really 
debian-specific, don't worry)

HTH.

Michal

Re: [Mysqlfs-general] Maintaining previous versions.

[Stef B. <st...@bo...>]

Michal Ludvig wrote:
> Binlog was designed for a totally different purpose and as such is 
> hardly usable for a convenient revision tracking on mysqlfs.
>
> For instance you'll have to keep all binlogs from when the FS was 
> created to be able to "replay" the filesystem to some point in time (and 
> honestly who keeps binlogs longer then a few days after they get their 
> way to slaves).
>   
Well, I did not know this way there are backups already available in the 
MySQL server.

I've just read the MySQL reference on the Internet, and it is possible 
to do binary logging
for at least the mysqlfs database.
> It would work on the whole-filesystem level, not for particular files 
> (which may or may not be what Stef wanted).
>   

Yes, indeed. With mysqlbinlog it's possible to go back to a previous 
version of a database, and copy that one to
a temporary database.  But this is step one. The next step would be to 
find the particular file (for a given time).

Ok, it's possible this way, but very complicated if you would like to 
check only one file.

But when reacting on your first remark. In my idea I woul like to use 
the mysql filesystem as a backup, not as
a filesystem where users have direct access to. I do understand that 
when users (or processes) have direct access, there will a lot of 
changes, which all get maintained...

My idea is to only do a copy or a sync of a directory of my choice, for 
example my homedirectory, once a day (4.30 at night for example), to the 
mounted mysqlfs.
This way not every change will get his way to the database.

Of course there should be a policy on how long a version will stay in 
the database, and there should be some utilities to find previous 
versions. But that's not the most difficult step I think.

Stef Bon

Re: [Mysqlfs-general] Maintaining previous versions.

[Michal L. <mi...@lo...>]

Smart Weblications GmbH, Florian Wiessner wrote:
> Hi,
> 
> Michal Ludvig schrieb:
>> But that all depends on your exact requirements.
> 
> Yes, you are right. And for that purpose there is mysql-binlog.

Binlog was designed for a totally different purpose and as such is 
hardly usable for a convenient revision tracking on mysqlfs.

For instance you'll have to keep all binlogs from when the FS was 
created to be able to "replay" the filesystem to some point in time (and 
honestly who keeps binlogs longer then a few days after they get their 
way to slaves).

It would work on the whole-filesystem level, not for particular files 
(which may or may not be what Stef wanted).

It isn't user-accessible.

Etc.

Michal

Re: [Mysqlfs-general] Maintaining previous versions.

[Smart W. G. F. W. <f.w...@sm...>]

Hi,

Michal Ludvig schrieb:
> Hi,
> 
> - I personally think you'll run out of space pretty quickly ;-)
> 
> - There need to be some way to remove old revisions.
> 
> - You probably only want to keep revisions of files in some defined 
> state, i.e. when they are closed. Keeping each and every revision after 
> each write() is useless as in most cases the file won't be consistent 
> anyway.
> 
> But that all depends on your exact requirements.

Yes, you are right. And for that purpose there is mysql-binlog.


-- 

Mit freundlichen Grüßen,

Florian Wiessner

Smart Weblications GmbH
Martinsberger Str. 1
D-95119 Naila

fon.: +49 700 762 789 32 - 0,12 EUR/Min*
fax.: +49 700 762 789 32 - 0,12 EUR/Min*
24/7: +49 900 311 886 00 - 1,99 EUR/Min*
http://www.smart-weblications.de

--
Sitz der Gesellschaft: Naila
Geschäftsführer: Florian Wiessner
HRB-Nr.: HRB 3840 Amtsgericht Hof
*aus dem dt. Festnetz, ggf. abweichende Preise aus dem Mobilfunknetz

Re: [Mysqlfs-general] Maintaining previous versions.

[Michal L. <mi...@lo...>]

Hi,

- I personally think you'll run out of space pretty quickly ;-)

- There need to be some way to remove old revisions.

- You probably only want to keep revisions of files in some defined 
state, i.e. when they are closed. Keeping each and every revision after 
each write() is useless as in most cases the file won't be consistent 
anyway.

But that all depends on your exact requirements.

Michal

Stef Bon wrote:
> Hello,
> 
> I would like to try to add the ability to mysqlfs to keep track of all 
> previous versions
> of files and directories (and..). Imho mysqlfs is ver suitable for doing 
> so, because it's already in a database.
> 
> First (before getting technical) I would like to know how you (reader) 
> feel about this. Is there anyone having thoughts
> about this?
> 
> Stef Bon
> 
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> _______________________________________________
> Mysqlfs-general mailing list
> Mys...@li...
> https://lists.sourceforge.net/lists/listinfo/mysqlfs-general

[Mysqlfs-general] Maintaining previous versions.

[Stef B. <st...@bo...>]

Hello,

I would like to try to add the ability to mysqlfs to keep track of all 
previous versions
of files and directories (and..). Imho mysqlfs is ver suitable for doing 
so, because it's already in a database.

First (before getting technical) I would like to know how you (reader) 
feel about this. Is there anyone having thoughts
about this?

Stef Bon

Re: [Mysqlfs-general] query_inode_full: how does it work?

[Stef B. <st...@bo...>]

Stef Bon wrote:
> Hello,
>
> now I can check the flow of what happens (by activating debugging). It's 
> much more clear now.
>
> But when I look at query_inode_full, I do not understand what happens 
> there at all.
> The purpose of this function is to translate a path into a inode, right? 
> But it does much more things.
> It "returns" a table nlinks, the inode, the parent. (and more?)
>
> Can someone explain why this is? It runs complicated queries....
>   

Ok,

now I've read it a couple of times, I understand it more or less.

A (long) query is constructed to determine the inode, with the help of a 
LEFT JOIN.

This query does grow bigger as the path is longer (deeper).

I'm looking into the mysql syntax now. Any hints about it welcome.

Stef Bon

[Mysqlfs-general] query_inode_full: how does it work?

[Stef B. <st...@bo...>]

Hello,

now I can check the flow of what happens (by activating debugging). It's 
much more clear now.

But when I look at query_inode_full, I do not understand what happens 
there at all.
The purpose of this function is to translate a path into a inode, right? 
But it does much more things.
It "returns" a table nlinks, the inode, the parent. (and more?)

Can someone explain why this is? It runs complicated queries....

Thanks in advance,

Stef Bon

Re: [Mysqlfs-general] Some questions.

[Michal L. <mi...@lo...>]

Stef Bon wrote:
> Michal Ludvig wrote:
>> have a look at mysqlfs_truncate() in mysqlfs.c and query_truncate() in
>> query.c. These two ensure that the non-used blocks are removed.
> 
> Ok I've found these functions. Are they called by Fuse, cause I did not 
> find where this truncate
> will be done.

Yes, mysqlfs_truncate is called by fuse, or more precisely, by the
kernel. Recompile with debugging on and you'll get a log message for
each function called. Then you will get a picture of the workflow.

Michal

Re: [Mysqlfs-general] Some questions.

[Stef B. <st...@bo...>]

Michal Ludvig wrote:
> Stef Bon wrote:
>
>   
>> First the overwriting of a file by another smaller one. This makes
>> (if both files differ much in size and both are big enough) that 
>> datablocks will be deleted. I've been searching the
>> sourcecode, but cannot find out how this works exactly. So before a file 
>> is written, there should be a procedure which
>> checks the file already exits, and removes the datablocks which will not 
>> be used. Is this right? If so, where can I find this
>> in the sourcecode?
>>     
>
> Hi,
> have a look at mysqlfs_truncate() in mysqlfs.c and query_truncate() in
> query.c. These two ensure that the non-used blocks are removed.
>
>   
Thanks for the fast reply.

Ok I've found these functions. Are they called by Fuse, cause I did not 
find where this truncate
will be done.

Stef Bon
> Michal
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Mysqlfs-general mailing list
> Mys...@li...
> https://lists.sourceforge.net/lists/listinfo/mysqlfs-general
>
>   

Re: [Mysqlfs-general] Some questions.

[Michal L. <mi...@lo...>]

Stef Bon wrote:

> First the overwriting of a file by another smaller one. This makes
> (if both files differ much in size and both are big enough) that 
> datablocks will be deleted. I've been searching the
> sourcecode, but cannot find out how this works exactly. So before a file 
> is written, there should be a procedure which
> checks the file already exits, and removes the datablocks which will not 
> be used. Is this right? If so, where can I find this
> in the sourcecode?

Hi,
have a look at mysqlfs_truncate() in mysqlfs.c and query_truncate() in
query.c. These two ensure that the non-used blocks are removed.

Michal

[Mysqlfs-general] Some questions.

[Stef B. <st...@bo...>]

Hello,

earlier I've posted a message in this list.
I've mentioned the "multithreading (im)possibilities" of mysqlfs. I thought
mysqlfs does not support multithreading. Just recent I've discovered that it
does. At least, it's possible to make use of the same mountpoint by more 
users
the same time.

I would like to investigate it's possible to add support for 
"maintaining versions of files and directories".
I think this is very possible because it's in a database. But first I'm 
trying to understand how mysqlfs
works. So I've got some questions. If anyone would be so kind to answer 
them.

First the overwriting of a file by another smaller one. This makes
(if both files differ much in size and both are big enough) that 
datablocks will be deleted. I've been searching the
sourcecode, but cannot find out how this works exactly. So before a file 
is written, there should be a procedure which
checks the file already exits, and removes the datablocks which will not 
be used. Is this right? If so, where can I find this
in the sourcecode?


Stef Bon