Mount and file system size issue

Mount and file system size issue

[bo]

Hi Reisers,

I do not understand well about "Reiserfs can be run on any linux supported
arch with
Linux2.4.13 and above." from FAQ of www.namesys.com.

I am running the ARM-Linux, which is based on Debian 2.4.5.

I tried to make and mount the reiserfs on 20G size, it was OK.
However, for 120G size, it is OK to make the file system but
I can not mount it.

It is really strange. Sometimes it is mounted after 1 hour of "mount"
command.
I could see it by "df".

Most of time it never come back from "mount -reiserfs ".

Is it really a problem of Linux revision?
Should I upgrade my linux to 2.4.13 or above to make the REISERFS work?

Thanks in advance,

Bo

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Thu, Jun 06, 2002 at 02:09:31PM -0700, bo wrote:

> I am running the ARM-Linux, which is based on Debian 2.4.5.

2.4.5 is way too outdated for purposes of using reiserfs.

> I tried to make and mount the reiserfs on 20G size, it was OK.
> However, for 120G size, it is OK to make the file system but
> I can not mount it.

What do you mean by "cannot mount"?
What is the error message if you get one?

> Most of time it never come back from "mount -reiserfs ".

Ah, so it just mounts, but does very slow.
BTW, how you get 120GB on your ARM system?
Do you have IDE controller and HDD attached to it or is this something
accessed over network?

> Is it really a problem of Linux revision?
> Should I upgrade my linux to 2.4.13 or above to make the REISERFS work?

Upgrading to 2.4.18 is highly advisable.

Bye,
    Oleg

Re: Mount and file system size issue

[bo]

It is our new product.
When I use ext2fs it mounts 120G filesystem successfully in about 3-4
seconds,
but I waited for about an hour with the REISERFS, never came back.

I know Linux 2.4.5 is far behind from the release. I want to know the reason
why
I have this MOUNTproblem for the big FS.

Thanks,

Bo
----- Original Message -----
From: "Oleg Drokin" <green@namesys.com>
To: "bo" <bo@sosnetwork.net>
Cc: <reiserfs-list@namesys.com>
Sent: Friday, June 07, 2002 12:21 AM
Subject: Re: [reiserfs-list] Mount and file system size issue


> Hello!
>
> On Thu, Jun 06, 2002 at 02:09:31PM -0700, bo wrote:
>
> > I am running the ARM-Linux, which is based on Debian 2.4.5.
>
> 2.4.5 is way too outdated for purposes of using reiserfs.
>
> > I tried to make and mount the reiserfs on 20G size, it was OK.
> > However, for 120G size, it is OK to make the file system but
> > I can not mount it.
>
> What do you mean by "cannot mount"?
> What is the error message if you get one?
>
> > Most of time it never come back from "mount -reiserfs ".
>
> Ah, so it just mounts, but does very slow.
> BTW, how you get 120GB on your ARM system?
> Do you have IDE controller and HDD attached to it or is this something
> accessed over network?
>
> > Is it really a problem of Linux revision?
> > Should I upgrade my linux to 2.4.13 or above to make the REISERFS work?
>
> Upgrading to 2.4.18 is highly advisable.
>
> Bye,
>     Oleg
>

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

   reiserfs reading all of the bitmaps immediately and stick bitmaps in the RAM.
   So if reading all of the bitmaps on 120G volume, that means
   reading 960 4k sectors off the media (each 32M there is bitmap on disk).
   That would eat at least 4 megabytes of RAM.
   Actually if you can do a profiling to see what reiserfs does during mounting,
   that would be interesting.
   Also you have not answered what kind of media you are using.

   ext2/3 read bitmap blocks on demand on the other side.

Bye,
    Oleg
On Fri, Jun 07, 2002 at 08:49:13AM -0700, bo wrote:
> It is our new product.
> When I use ext2fs it mounts 120G filesystem successfully in about 3-4
> seconds,
> but I waited for about an hour with the REISERFS, never came back.
> 
> I know Linux 2.4.5 is far behind from the release. I want to know the reason
> why
> I have this MOUNTproblem for the big FS.
> 
> Thanks,
> 
> Bo
> ----- Original Message -----
> From: "Oleg Drokin" <green@namesys.com>
> To: "bo" <bo@sosnetwork.net>
> Cc: <reiserfs-list@namesys.com>
> Sent: Friday, June 07, 2002 12:21 AM
> Subject: Re: [reiserfs-list] Mount and file system size issue
> 
> 
> > Hello!
> >
> > On Thu, Jun 06, 2002 at 02:09:31PM -0700, bo wrote:
> >
> > > I am running the ARM-Linux, which is based on Debian 2.4.5.
> >
> > 2.4.5 is way too outdated for purposes of using reiserfs.
> >
> > > I tried to make and mount the reiserfs on 20G size, it was OK.
> > > However, for 120G size, it is OK to make the file system but
> > > I can not mount it.
> >
> > What do you mean by "cannot mount"?
> > What is the error message if you get one?
> >
> > > Most of time it never come back from "mount -reiserfs ".
> >
> > Ah, so it just mounts, but does very slow.
> > BTW, how you get 120GB on your ARM system?
> > Do you have IDE controller and HDD attached to it or is this something
> > accessed over network?
> >
> > > Is it really a problem of Linux revision?
> > > Should I upgrade my linux to 2.4.13 or above to make the REISERFS work?
> >
> > Upgrading to 2.4.18 is highly advisable.
> >
> > Bye,
> >     Oleg
> >
> 

Re: Mount and file system size issue

[bo]

Oleg,

 I have several IDE controllers on the board.
My system have 65Mbytes RAM on the board, and it uses
256 Mbytes as swap devices.

I tried the same on the PC-386 linux, it comes in about 2-3 seconds for 120G
Reiserfs.
I use linux2.4.18 on my PC.

According to your openion and FAQ on www.namesys.com , there may be two
issues;
one is RAM shortage for bitmap, another is to use Linux2.4.13 or above for
ARM platform.

Could it be both?

Could you help me what( or how) I need to investigate?

Thanks,

Bo

----- Original Message -----
From: "Oleg Drokin" <green@namesys.com>
To: "bo" <bo@sosnetwork.net>
Cc: <reiserfs-list@namesys.com>
Sent: Friday, June 07, 2002 9:02 AM
Subject: Re: [reiserfs-list] Mount and file system size issue


> Hello!
>
>    reiserfs reading all of the bitmaps immediately and stick bitmaps in
the RAM.
>    So if reading all of the bitmaps on 120G volume, that means
>    reading 960 4k sectors off the media (each 32M there is bitmap on
disk).
>    That would eat at least 4 megabytes of RAM.
>    Actually if you can do a profiling to see what reiserfs does during
mounting,
>    that would be interesting.
>    Also you have not answered what kind of media you are using.
>
>    ext2/3 read bitmap blocks on demand on the other side.
>
> Bye,
>     Oleg
> On Fri, Jun 07, 2002 at 08:49:13AM -0700, bo wrote:
> > It is our new product.
> > When I use ext2fs it mounts 120G filesystem successfully in about 3-4
> > seconds,
> > but I waited for about an hour with the REISERFS, never came back.
> >
> > I know Linux 2.4.5 is far behind from the release. I want to know the
reason
> > why
> > I have this MOUNTproblem for the big FS.
> >
> > Thanks,
> >
> > Bo
> > ----- Original Message -----
> > From: "Oleg Drokin" <green@namesys.com>
> > To: "bo" <bo@sosnetwork.net>
> > Cc: <reiserfs-list@namesys.com>
> > Sent: Friday, June 07, 2002 12:21 AM
> > Subject: Re: [reiserfs-list] Mount and file system size issue
> >
> >
> > > Hello!
> > >
> > > On Thu, Jun 06, 2002 at 02:09:31PM -0700, bo wrote:
> > >
> > > > I am running the ARM-Linux, which is based on Debian 2.4.5.
> > >
> > > 2.4.5 is way too outdated for purposes of using reiserfs.
> > >
> > > > I tried to make and mount the reiserfs on 20G size, it was OK.
> > > > However, for 120G size, it is OK to make the file system but
> > > > I can not mount it.
> > >
> > > What do you mean by "cannot mount"?
> > > What is the error message if you get one?
> > >
> > > > Most of time it never come back from "mount -reiserfs ".
> > >
> > > Ah, so it just mounts, but does very slow.
> > > BTW, how you get 120GB on your ARM system?
> > > Do you have IDE controller and HDD attached to it or is this something
> > > accessed over network?
> > >
> > > > Is it really a problem of Linux revision?
> > > > Should I upgrade my linux to 2.4.13 or above to make the REISERFS
work?
> > >
> > > Upgrading to 2.4.18 is highly advisable.
> > >
> > > Bye,
> > >     Oleg
> > >
> >
>

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Fri, Jun 07, 2002 at 09:38:01AM -0700, bo wrote:

> one is RAM shortage for bitmap, another is to use Linux2.4.13 or above for
> ARM platform.
> Could it be both?

It is unclear.

> Could you help me what( or how) I need to investigate?

First try 2.4.18, if the symptoms are the same, try to profile a kernel
(man readprofile).

BTW, what kind of ARM system do you have?

Bye,
    Oleg

Re: Mount and file system size issue

[bo]

Hello,

I have more information on this issue as folow;

System:  ARM9, 32M RAM, IDE, debian_arm_linux2.4.5
filesystem size = 120 G
When to happen:

   > mkreiserfs /dev/hdf1
   or
   >mount -t reiserfs /dev/hdf1 /mnt/hdf1

  The cosole message:

   >reiserfs: warning: CONFIG_REISERFS_CHECK is set On
   >reiserfs: warning:  -it is slow mode for debugging
   >reiserfs:  checking transaction log(device 21:01)

???   It took 2-3 seconds(1 out of 5), or 3-4 munites(1 out of 5)
        or forever(3 out of 5) to complete.

System status at this time: from another console( run "top")
---------------------------------------------------------
  1:40am  up  1:40,  2 users,  load average: 8.62, 7.71, 4.81
37 processes: 30 sleeping, 6 running, 1 zombie, 0 stopped
CPU states:  0.3% user, 99.6% system,  0.0% nice,  0.0% idle
Mem:   30020K av,  28852K used,   1168K free,      0K shrd,  14132K buff
Swap: 262576K av,   4204K used, 258372K free                  4160K cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME COMMAND
    3 root      16   0     0    0     0 RW      0 34.0  0.0   7:16 kswapd
 1283 root      17   0  4956 4956   520 R       0 13.1 16.5   0:56
mkreiserfs
  328 jskim     15   0   612  536   536 R       0 11.5  1.7   6:50 top
----------------------------------------

Any sugestions?

Thanks,

Bo




----- Original Message -----
From: "Oleg Drokin" <green@namesys.com>
To: "bo" <bo@sosnetwork.net>
Cc: <reiserfs-list@namesys.com>
Sent: Friday, June 07, 2002 12:26 PM
Subject: Re: [reiserfs-list] Mount and file system size issue


> Hello!
>
> On Fri, Jun 07, 2002 at 09:38:01AM -0700, bo wrote:
>
> > one is RAM shortage for bitmap, another is to use Linux2.4.13 or above
for
> > ARM platform.
> > Could it be both?
>
> It is unclear.
>
> > Could you help me what( or how) I need to investigate?
>
> First try 2.4.18, if the symptoms are the same, try to profile a kernel
> (man readprofile).
>
> BTW, what kind of ARM system do you have?
>
> Bye,
>     Oleg
>

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Sat, Jun 08, 2002 at 12:42:21PM -0700, bo wrote:

>    >reiserfs: warning: CONFIG_REISERFS_CHECK is set On
>    >reiserfs: warning:  -it is slow mode for debugging
>    >reiserfs:  checking transaction log(device 21:01)

Hm. Would it help if you turn of CONFIG_REISERFS_CHECK config item?

> ???   It took 2-3 seconds(1 out of 5), or 3-4 munites(1 out of 5)
>         or forever(3 out of 5) to complete.

All of that on the same filesystem (with same contents)?

> System status at this time: from another console( run "top")
> ---------------------------------------------------------
>   1:40am  up  1:40,  2 users,  load average: 8.62, 7.71, 4.81
> 37 processes: 30 sleeping, 6 running, 1 zombie, 0 stopped
> CPU states:  0.3% user, 99.6% system,  0.0% nice,  0.0% idle

Ok, all the time is spent in kernel.
Can you capture several traces for us please?
( several sysrq-p and/or sysrq-t outputs when this happens).

> Swap: 262576K av,   4204K used, 258372K free                  4160K cached
>   PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME COMMAND
>     3 root      16   0     0    0     0 RW      0 34.0  0.0   7:16 kswapd
>  1283 root      17   0  4956 4956   520 R       0 13.1 16.5   0:56
> mkreiserfs

Hm. Mkreiserfs? I though you expirience the promlems when you
try to mount 120G filesystem, no?

Bye,
    Olet

Re: Mount and file system size issue

[bo]

Hi ,

> On Sat, Jun 08, 2002 at 12:42:21PM -0700, bo wrote:
>
> >    >reiserfs: warning: CONFIG_REISERFS_CHECK is set On
> >    >reiserfs: warning:  -it is slow mode for debugging
> >    >reiserfs:  checking transaction log(device 21:01)
>
> Hm. Would it help if you turn of CONFIG_REISERFS_CHECK config item?
>

I will try it later, I could not rebuild the system now.
Are there anyway to turn it off on the command line?

> > ???   It took 2-3 seconds(1 out of 5), or 3-4 munites(1 out of 5)
> >         or forever(3 out of 5) to complete.
>
> All of that on the same filesystem (with same contents)?
>

I mean it fails(hang) most of time when I try to mount the reiserfs with
120G size after "mkreiserfs".
As you notice in the system status below, the CPU is occupied by the system
by 99.6(8) %.
I do not know why?

> > System status at this time: from another console( run "top")
> > ---------------------------------------------------------
> >   1:40am  up  1:40,  2 users,  load average: 8.62, 7.71, 4.81
> > 37 processes: 30 sleeping, 6 running, 1 zombie, 0 stopped
> > CPU states:  0.3% user, 99.6% system,  0.0% nice,  0.0% idle
>
> Ok, all the time is spent in kernel.
> Can you capture several traces for us please?
> ( several sysrq-p and/or sysrq-t outputs when this happens).
>

Sorry! What is "sysrq"?

> > Swap: 262576K av,   4204K used, 258372K free                  4160K
cached
> >   PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME
COMMAND
> >     3 root      16   0     0    0     0 RW      0 34.0  0.0   7:16
kswapd
> >  1283 root      17   0  4956 4956   520 R       0 13.1 16.5   0:56
> > mkreiserfs
>
> Hm. Mkreiserfs? I though you expirience the promlems when you
> try to mount 120G filesystem, no?
>

Sorry I misled you!
This problem happens most of time at "mount", but it happens sometimes at
"mkreiserfs".

One more strange thing is "failed to generate UUID" as follow;

-------------
root@praetorian:/# mkreiserfs  /dev/hdg1

<-------------mkreiserfs, 2002------------->
reiserfsprogs 3.x.1b

mkreiserfs: Guessing about desired format..
mkreiserfs: Kernel 2.4.5-rmk4 is running.
failed to genetate UUID
<----------------------Why????
Format 3.6 with standard journal
Count of blocks on the device: 30015208
Number of blocks consumed by mkreiserfs formatting process: 9127
Blocksize: 4096
Hash function used to sort names: "r5"
Journal Size 8193 blocks (first block 18)
Journal Max transaction length 1024
inode generation number: 0
UUID: 00000000-0000-0000-0000-000000000000
------------------------

Do you have any clue from it?

Thanks,

Bo

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Mon, Jun 10, 2002 at 10:47:00AM -0700, bo wrote:
> > >    >reiserfs: warning:  -it is slow mode for debugging
> > >    >reiserfs:  checking transaction log(device 21:01)
> > Hm. Would it help if you turn of CONFIG_REISERFS_CHECK config item?
> I will try it later, I could not rebuild the system now.
> Are there anyway to turn it off on the command line?

No, this cannot be turned off on command line.

> > > System status at this time: from another console( run "top")
> > > ---------------------------------------------------------
> > >   1:40am  up  1:40,  2 users,  load average: 8.62, 7.71, 4.81
> > > 37 processes: 30 sleeping, 6 running, 1 zombie, 0 stopped
> > > CPU states:  0.3% user, 99.6% system,  0.0% nice,  0.0% idle
> > Ok, all the time is spent in kernel.
> > Can you capture several traces for us please?
> > ( several sysrq-p and/or sysrq-t outputs when this happens).
> Sorry! What is "sysrq"?

See Documentation/sysrq.txt in your linux kernel source.

> > >   PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME
> COMMAND
> > >     3 root      16   0     0    0     0 RW      0 34.0  0.0   7:16
> kswapd
> > >  1283 root      17   0  4956 4956   520 R       0 13.1 16.5   0:56
> > > mkreiserfs
> > Hm. Mkreiserfs? I though you expirience the promlems when you
> > try to mount 120G filesystem, no?
> Sorry I misled you!
> This problem happens most of time at "mount", but it happens sometimes at
> "mkreiserfs".

Strange. If mkreiserfs hangs, it may be that your block device driver
is unhappy.

> One more strange thing is "failed to generate UUID" as follow;

Do you have /dev/urandom?

Bye,
    Oleg

Re: Mount and file system size issue

[Kuba Ober]

On poniedzia³ek 10 czerwiec 2002 01:47 pm, bo wrote:
> Hi ,
>
> > On Sat, Jun 08, 2002 at 12:42:21PM -0700, bo wrote:
> > >    >reiserfs: warning: CONFIG_REISERFS_CHECK is set On
> > >    >reiserfs: warning:  -it is slow mode for debugging
> > >    >reiserfs:  checking transaction log(device 21:01)
> >
> > Hm. Would it help if you turn of CONFIG_REISERFS_CHECK config item?
>
> I will try it later, I could not rebuild the system now.
> Are there anyway to turn it off on the command line?
>
> > > ???   It took 2-3 seconds(1 out of 5), or 3-4 munites(1 out of 5)
> > >         or forever(3 out of 5) to complete.
> >
> > All of that on the same filesystem (with same contents)?
>
> I mean it fails(hang) most of time when I try to mount the reiserfs with
> 120G size after "mkreiserfs".
> As you notice in the system status below, the CPU is occupied by the system
> by 99.6(8) %.
> I do not know why?
>
> > > System status at this time: from another console( run "top")
> > > ---------------------------------------------------------
> > >   1:40am  up  1:40,  2 users,  load average: 8.62, 7.71, 4.81
> > > 37 processes: 30 sleeping, 6 running, 1 zombie, 0 stopped
> > > CPU states:  0.3% user, 99.6% system,  0.0% nice,  0.0% idle
> >
> > Ok, all the time is spent in kernel.
> > Can you capture several traces for us please?
> > ( several sysrq-p and/or sysrq-t outputs when this happens).
>
> Sorry! What is "sysrq"?

It's a key on your keyboard, I hope.

# echo 1 > /proc/sys/kernel sysrq
# cat /usr/src/linux/Documentation/sysrq.txt

Cheers, Kuba Ober

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Wed, Jun 12, 2002 at 11:33:31AM -0400, Kuba Ober wrote:

> > Sorry! What is "sysrq"?
> It's a key on your keyboard, I hope.

I afraid that most probably his system does not have keyboard and
operated through serial console of some kind.
Of course I might be wrong, but it seems he haev some kind of
embedded board (why do they use ARM CPU otherwise?)

Bye,
    Oleg

Re: Mount and file system size issue

[Kuba Ober]

On ¶roda 12 czerwiec 2002 02:38 pm, you wrote:
> Hello!
>
> On Wed, Jun 12, 2002 at 11:33:31AM -0400, Kuba Ober wrote:
> > > Sorry! What is "sysrq"?
> >
> > It's a key on your keyboard, I hope.
>
> I afraid that most probably his system does not have keyboard and
> operated through serial console of some kind.
> Of course I might be wrong, but it seems he haev some kind of
> embedded board (why do they use ARM CPU otherwise?)

Hmmm... I would have to look into the sources, but wouldn't the sysrq have 
some encoding even through serial console (it's console after all)...

Cheers, Kuba

Re: Mount and file system size issue

[bo]

Hi,

You are right.
My system does not have any I/O devices such as mouse or KBD.

Of course you can connected to the system through ethernet(telnet).

I am investigating the solution from "sysrq.txt".

Thanks,

Bo
----- Original Message ----- 
From: "Oleg Drokin" <green@namesys.com>
To: "Kuba Ober" <kuba@mareimbrium.org>
Cc: "bo" <bo@sosnetwork.net>; <reiserfs-list@namesys.com>
Sent: Wednesday, June 12, 2002 11:38 AM
Subject: Re: [reiserfs-list] Mount and file system size issue


> Hello!
> 
> On Wed, Jun 12, 2002 at 11:33:31AM -0400, Kuba Ober wrote:
> 
> > > Sorry! What is "sysrq"?
> > It's a key on your keyboard, I hope.
> 
> I afraid that most probably his system does not have keyboard and
> operated through serial console of some kind.
> Of course I might be wrong, but it seems he haev some kind of
> embedded board (why do they use ARM CPU otherwise?)
> 
> Bye,
>     Oleg
> 

Re: Mount and file system size issue

[bscott]

On Wed, 12 Jun 2002, at 2:55pm, Kuba Ober wrote:
> Hmmm... I would have to look into the sources, but wouldn't the sysrq have
> some encoding even through serial console (it's console after all)...

  It does.  Send a serial BREAK signal, followed by the command key letter.

-- 
Ben Scott <bscott@ntisys.com>
| The opinions expressed in this message are those of the author and do not |
| necessarily represent the views or policy of any other person, entity or  |
| organization.  All information is provided without warranty of any kind.  |

Re: Mount and file system size issue

[bmoon]

Reisers,

I have three questions related with this 4 months old issue.

- Still reiser2 or reiser4 do not support "on demand" bitmap read?
- What is the supporting block size on reiser2, and reiser4?
- I am going to build one big volume(LVM) out of 8 IDE disks and make
reiserfs on it.
  Where does the system write the meta data(super block, bitmap,..) into?
Will it copy the same meta data on each disk?

Thanks,

Bo
----- Original Message -----
From: "Oleg Drokin" <green@namesys.com>
Subject: Re: [reiserfs-list] Mount and file system size issue


> Hello!
>
>    reiserfs reading all of the bitmaps immediately and stick bitmaps in
the RAM.
>    So if reading all of the bitmaps on 120G volume, that means
>    reading 960 4k sectors off the media (each 32M there is bitmap on
disk).
>    That would eat at least 4 megabytes of RAM.
>    Actually if you can do a profiling to see what reiserfs does during
mounting,
>    that would be interesting.
>    Also you have not answered what kind of media you are using.
>
>    ext2/3 read bitmap blocks on demand on the other side.
>
>

Re: Mount and file system size issue

[Oleg Drokin]

Hello!

On Thu, Nov 07, 2002 at 04:30:17PM -0800, bmoon wrote:

> I have three questions related with this 4 months old issue.
> - Still reiser2 or reiser4 do not support "on demand" bitmap read?

That's true for reiser3. reiser4 have on-demand bitmap reading,
but not unloading until umount as I remember.

> - What is the supporting block size on reiser2, and reiser4?

Right now due for simplicity reiser4 blocksize should be equal to
PAGE_CACHE_SIZE on underlying architecture.
reiser3 have 4096 bytes blocksize by default (this is the only supportted
blocksize for 2.4 kernel series). Other possible values are 1024, 2048, ...
and so on until you hit PAGE_CACHE_SIZE limit (Linux limit 4096 for x86,
8128 for Alpha, variable for ia64)

> - I am going to build one big volume(LVM) out of 8 IDE disks and make
> reiserfs on it.
>   Where does the system write the meta data(super block, bitmap,..) into?

Superblock will be at the beginning of volume and will probably be mapped to
first disk. Bitmaps are spread over the disk (the ferquency of bitmaps
on disk depends on blocksize).

> Will it copy the same meta data on each disk?

No, unless you use RAID1 stuff, and even then this copying will be done
not by reiserfs itself.

Bye,
    Oleg