2 minutes for a reiser4 filesystem to mount.

2 minutes for a reiser4 filesystem to mount.

[Georgios Tsalikis]

The filesystem is 1.7GB large, lies in a GPT patition and is formated 
with -o node=node41 without any other override. The kernel is 4.2.3 with 
the 4.2.2 patch.
I tried the same on a 250MB partition and it mounts instantly. dmesg 
shows nothing special
Any ideas? Thanks in advance. :)

Re: 2 minutes for a reiser4 filesystem to mount.

[Jose R R]

Niltze!

On Thu, Oct 22, 2015 at 10:38 AM, Georgios Tsalikis
<georgios@tsalikis.net> wrote:
> The filesystem is 1.7GB large, lies in a GPT patition and is formated with
> -o node=node41
In this test machine(s) my partitions are not GPT

> without any other override. The kernel is 4.2.3 with the
> 4.2.2 patch.
I am running similar combination; mount time of ~60GB & ~50GB
partitions (at least one previously formatted: -o create=reg40) with
*existing data*  is under 8 seconds.

> I tried the same on a 250MB partition and it mounts instantly. dmesg shows
> nothing special
> Any ideas?

Tried formatting a ~24GB non-GPT partition with options as you
specified; the mount time was under 3 seconds for the newly formatted
partition:
https://pbs.twimg.com/media/CR-zkQtVAAA5D_G.png:large

I have no idea if GPT makes a difference.
> Thanks in advance. :)
>
> --
> To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



-- 
Jose R R
http://metztli.it
---------------------------------------------------------------------------------------------
Try at no charge http://b2evolution.net for http://OpenShift.com PaaS
---------------------------------------------------------------------------------------------
from our GitHub http://Nepohualtzintzin.com repository. Cloud the easy way!

Re: 2 minutes for a reiser4 filesystem to mount.

[Georgios Tsalikis]

Hello :)

> Niltze!
>
> On Thu, Oct 22, 2015 at 10:38 AM, Georgios Tsalikis
> <georgios@tsalikis.net> wrote:
>> The filesystem is 1.7GB large, lies in a GPT patition and is formated with
>> -o node=node41
> In this test machine(s) my partitions are not GPT
>
>> without any other override. The kernel is 4.2.3 with the
>> 4.2.2 patch.
> I am running similar combination; mount time of ~60GB & ~50GB
> partitions (at least one previously formatted: -o create=reg40) with
> *existing data*  is under 8 seconds.
>
>> I tried the same on a 250MB partition and it mounts instantly. dmesg shows
>> nothing special
>> Any ideas?
> Tried formatting a ~24GB non-GPT partition with options as you
> specified; the mount time was under 3 seconds for the newly formatted
> partition:
> https://pbs.twimg.com/media/CR-zkQtVAAA5D_G.png:large

> I noticed that yours uses 4096K block size. My partition uses 64K. Is this
> due to the much larger size of my partition? Maybe it is the problem behind the delay?
> I have no idea if GPT makes a difference.
>> Thanks in advance. :)
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>

Re: 2 minutes for a reiser4 filesystem to mount.

[Morgan Smith]

I had this behavior a couple years in the past and in my case this was
due to some metadata being sought and read at mount time. There was a
mount option to disable that behavior though I haven't used it in a long
time and cannot remember the option however Q19 this gentoo forum
reiser4 faq [1] suggests that it was the dont_load_bitmap mount option.
I do not know if such an option is still available for the current version.

[1] https://forums.gentoo.org/viewtopic-t-706171.html

On 10/22/2015 11:38 AM, Georgios Tsalikis wrote:
> The filesystem is 1.7GB large, lies in a GPT patition and is formated
> with -o node=node41 without any other override. The kernel is 4.2.3 with
> the 4.2.2 patch.
> I tried the same on a 250MB partition and it mounts instantly. dmesg
> shows nothing special
> Any ideas? Thanks in advance. :)
> 
> -- 
> To unsubscribe from this list: send the line "unsubscribe
> reiserfs-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2 minutes for a reiser4 filesystem to mount.

[Ivan Shapovalov]

[-- Attachment #1: Type: text/plain, Size: 737 bytes --]

On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
> The filesystem is 1.7GB large, lies in a GPT patition and is formated
> with -o node=node41 without any other override. The kernel is 4.2.3
> with 
> the 4.2.2 patch.
> I tried the same on a 250MB partition and it mounts instantly. dmesg 
> shows nothing special
> Any ideas? Thanks in advance. :)

Hello,

try "dont_load_bitmap" in mount options. However, this is anyway very
strange for a ~2GB partition...

> 
> --
> To unsubscribe from this list: send the line "unsubscribe reiserfs-
> devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
Ivan Shapovalov / intelfx /


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 213 bytes --]

Re: 2 minutes for a reiser4 filesystem to mount.

[Georgios Tsalikis]

On 24/10/2015 12:37 πμ, Ivan Shapovalov wrote:
> On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
>> The filesystem is 1.7GB large, lies in a GPT patition and is formated
>> with -o node=node41 without any other override. The kernel is 4.2.3
>> with
>> the 4.2.2 patch.
>> I tried the same on a 250MB partition and it mounts instantly. dmesg
>> shows nothing special
>> Any ideas? Thanks in advance. :)
> Hello,
>
> try "dont_load_bitmap" in mount options. However, this is anyway very
> strange for a ~2GB partition...

I am sorry, did i write 1.7GB? I meant 1.7TB... i am old and such figures dont fit my mind easily!

>> --
>> To unsubscribe from this list: send the line "unsubscribe reiserfs-
>> devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2 minutes for a reiser4 filesystem to mount.

[Ivan Shapovalov]

> 24 окт. 2015 г., в 1:06, Georgios Tsalikis <georgios.tsalikis@hol.gr> написал(а):
> 
>> On 24/10/2015 12:37 πμ, Ivan Shapovalov wrote:
>>> On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
>>> The filesystem is 1.7GB large, lies in a GPT patition and is formated
>>> with -o node=node41 without any other override. The kernel is 4.2.3
>>> with
>>> the 4.2.2 patch.
>>> I tried the same on a 250MB partition and it mounts instantly. dmesg
>>> shows nothing special
>>> Any ideas? Thanks in advance. :)
>> Hello,
>> 
>> try "dont_load_bitmap" in mount options. However, this is anyway very
>> strange for a ~2GB partition...
> 
> I am sorry, did i write 1.7GB? I meant 1.7TB... i am old and such figures dont fit my mind easily!

Ah, so it's 1.7 TB? Then it's not unusual at all. Indeed, try mounting with `-o dont_load_bitmap`.

--
Ivan Shapovalov / intelfx /

(Sent from a phone. Havoc may be wreaked on the formatting.)--
To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2 minutes for a reiser4 filesystem to mount.

[Georgios Tsalikis]

On 24/10/2015 01:58 πμ, Ivan Shapovalov wrote:
>> 24 окт. 2015 г., в 1:06, Georgios Tsalikis <georgios.tsalikis@hol.gr> написал(а):
>>
>>> On 24/10/2015 12:37 πμ, Ivan Shapovalov wrote:
>>>> On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
>>>> The filesystem is 1.7GB large, lies in a GPT patition and is formated
>>>> with -o node=node41 without any other override. The kernel is 4.2.3
>>>> with
>>>> the 4.2.2 patch.
>>>> I tried the same on a 250MB partition and it mounts instantly. dmesg
>>>> shows nothing special
>>>> Any ideas? Thanks in advance. :)
>>> Hello,
>>>
>>> try "dont_load_bitmap" in mount options. However, this is anyway very
>>> strange for a ~2GB partition...
>> I am sorry, did i write 1.7GB? I meant 1.7TB... i am old and such figures dont fit my mind easily!
> Ah, so it's 1.7 TB? Then it's not unusual at all. Indeed, try mounting with `-o dont_load_bitmap`.
>
I did so and it worked. Thank you. But what does that option do? And which are the other options and their function?


--
To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2 minutes for a reiser4 filesystem to mount.

[Ivan Shapovalov]

[-- Attachment #1: Type: text/plain, Size: 2033 bytes --]

On 2015-10-24 at 22:41 +0300, Georgios Tsalikis wrote:
> On 24/10/2015 01:58 πμ, Ivan Shapovalov wrote:
> > > 24 окт. 2015 г., в 1:06, Georgios Tsalikis <georgios.tsalikis@hol
> > > .gr> написал(а):
> > > 
> > > > On 24/10/2015 12:37 πμ, Ivan Shapovalov wrote:
> > > > > On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
> > > > > The filesystem is 1.7GB large, lies in a GPT patition and is
> > > > > formated
> > > > > with -o node=node41 without any other override. The kernel is
> > > > > 4.2.3
> > > > > with
> > > > > the 4.2.2 patch.
> > > > > I tried the same on a 250MB partition and it mounts
> > > > > instantly. dmesg
> > > > > shows nothing special
> > > > > Any ideas? Thanks in advance. :)
> > > > Hello,
> > > > 
> > > > try "dont_load_bitmap" in mount options. However, this is
> > > > anyway very
> > > > strange for a ~2GB partition...
> > > I am sorry, did i write 1.7GB? I meant 1.7TB... i am old and such
> > > figures dont fit my mind easily!
> > Ah, so it's 1.7 TB? Then it's not unusual at all. Indeed, try
> > mounting with `-o dont_load_bitmap`.
> > 
> I did so and it worked. Thank you. But what does that option do? 

This option tells reiser4 to load bitmap blocks on demand, and not all
at once at filesystem mount time.

The bitmap is how reiser4 stores information about which on-disk blocks
are allocated and which are free (one bit corresponds to one block,
hence the name).

The 2TiB filesystem has 2*1024^4 / 4096 = 536870912 blocks, which take
536870912 / (8*4096) = 16384 bitmap blocks, so it takes 120/16384 ~ 7ms
to load one block, which is not that unusual (given that the bitmap
loading doesn't seem to be optimized too well: each block is read in a
separate request).

> And which are the other options and their function?

I'm afraid that the full list of reiser4 options can only be found in
code (file init_super.c, functions push_sb_field_opts() and
reiser4_init_super_data().

-- 
Ivan Shapovalov / intelfx /


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 213 bytes --]

Re: 2 minutes for a reiser4 filesystem to mount.

[Georgios Tsalikis]

On 25/10/2015 01:39 πμ, Ivan Shapovalov wrote:
> On 2015-10-24 at 22:41 +0300, Georgios Tsalikis wrote:
>> On 24/10/2015 01:58 πμ, Ivan Shapovalov wrote:
>>>> 24 окт. 2015 г., в 1:06, Georgios Tsalikis <georgios.tsalikis@hol
>>>> .gr> написал(а):
>>>>
>>>>> On 24/10/2015 12:37 πμ, Ivan Shapovalov wrote:
>>>>>> On 2015-10-22 at 20:38 +0300, Georgios Tsalikis wrote:
>>>>>> The filesystem is 1.7GB large, lies in a GPT patition and is
>>>>>> formated
>>>>>> with -o node=node41 without any other override. The kernel is
>>>>>> 4.2.3
>>>>>> with
>>>>>> the 4.2.2 patch.
>>>>>> I tried the same on a 250MB partition and it mounts
>>>>>> instantly. dmesg
>>>>>> shows nothing special
>>>>>> Any ideas? Thanks in advance. :)
>>>>> Hello,
>>>>>
>>>>> try "dont_load_bitmap" in mount options. However, this is
>>>>> anyway very
>>>>> strange for a ~2GB partition...
>>>> I am sorry, did i write 1.7GB? I meant 1.7TB... i am old and such
>>>> figures dont fit my mind easily!
>>> Ah, so it's 1.7 TB? Then it's not unusual at all. Indeed, try
>>> mounting with `-o dont_load_bitmap`.
>>>
>> I did so and it worked. Thank you. But what does that option do?
> This option tells reiser4 to load bitmap blocks on demand, and not all
> at once at filesystem mount time.
Does this mean using this option means a penalty? Maybe in some cases 
(servers etc) it is worth waiting the fs to mount?
>
> The bitmap is how reiser4 stores information about which on-disk blocks
> are allocated and which are free (one bit corresponds to one block,
> hence the name).
>
> The 2TiB filesystem has 2*1024^4 / 4096 = 536870912 blocks, which take
> 536870912 / (8*4096) = 16384 bitmap blocks, so it takes 120/16384 ~ 7ms
> to load one block, which is not that unusual (given that the bitmap
> loading doesn't seem to be optimized too well: each block is read in a
> separate request).
>
>> And which are the other options and their function?
> I'm afraid that the full list of reiser4 options can only be found in
> code (file init_super.c, functions push_sb_field_opts() and
> reiser4_init_super_data().
>

--
To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html