moosefs-users — Mailing list for MooseFS Users

[Moosefs-users] Deduplication

[Alexander A. <akh...@ri...>]

Hi all!

I wonder if you have plans to implement data deduplication in MooseFS ?
I mean searching for the equal chunks and make them
to be stored just once.
In my humble opinion it will be no so hard
with a glance each chunk have checksum already.

Thanks!
wbr
Alexander Akhobadze

Re: [Moosefs-users] problem: cannot access group directory

[Alexander A. <akh...@ri...>]

Hi all!

I have a problem like described is this post.

My config:
mfsexports.cfg on Master:
192.168.206.0/24                /       rw,alldirs,ignoregid,maproot=0

On client:
mfsmount /moosefs_mnt -H 192.168.206.1 -o mfscachefiles,suid
mkdir /moosefs_mnt/test
chmod g+s /moosefs_mnt/test

MFS Client is not used as MFS Master or Metalogger or Chunk server.
MFS Client is also a Win 2003 domain member Samba server
and directory /moosefs_mnt/test is shared by Samba.

chown "DOMAIN\some_group" /moosefs_mnt/test
ls -ld /moosefs_mnt/test
drwxrws--- 2 root some_group 0 2010..... /moosefs_mnt/test

Some domain-user is a member of "DOMAIN\some_group".
When he creates file or folder in /moosefs_mnt/test (via Samba share)
the newly created child folder does not inherit
groupID from its parent (as expected because the setgid flag is set):

ls -ld /moosefs_mnt/test/subdir
drwxrws--- 2 DOMAIN\domain-user NEWRIDAN\Domain users 0 2010..... /moosefs_mnt/test/subdir

As you can see newly created child folder has "Domain users"
group assigned. I think it is because "Domain users" is a primary group
for domain-user.

In the same time when test-folder is shared from EXT3 file system
setgid flag works perfectly.

Any idea? Thanks!
wbr
Alexander Akhobadze

======================================================

On Wed, 22 Sep 2010 11:41:00 +0200
Michał Borychowski <mic...@ge...> wrote:

> Please check this FAQ entry:
> http://www.moosefs.org/moosefs-faq.html#supplementary_groups

Oops, I missed the FAQ ... The entry says that MFS will test privileges only of main groups because it is much safer. I love "safer" so I would use the permission 755 instead of 750  

Thanks Michał for you helps,

> -----Original Message-----
> From: Anh K. Huynh [mailto:ky...@vi...] 
> Sent: Wednesday, September 22, 2010 10:52 AM
> To: moo...@li...
> Subject: [Moosefs-users] problem: cannot access group directory
> 
> Hello,
> 
> I have a user `bbt_landing` who belongs to the group `builder`, and
> a directory name `test` which has the permission 750. The directory
> is located on a MFS volume. The problem is that user `bbt_landing`
> cannot access to the directory.
> 
> More details can be found at
>       http://metakyanh.sarovar.org/moosefs.report/report3.html
> 
> Any idea?

Re: [Moosefs-users] problem: cannot access group directory

[Anh K. H. <ky...@vi...>]

On Wed, 22 Sep 2010 11:41:00 +0200
Michał Borychowski <mic...@ge...> wrote:

> Please check this FAQ entry:
> http://www.moosefs.org/moosefs-faq.html#supplementary_groups

Oops, I missed the FAQ ... The entry says that MFS will test privileges only of main groups because it is much safer. I love "safer" so I would use the permission 755 instead of 750 :P

Thanks Michał for you helps,

> -----Original Message-----
> From: Anh K. Huynh [mailto:ky...@vi...] 
> Sent: Wednesday, September 22, 2010 10:52 AM
> To: moo...@li...
> Subject: [Moosefs-users] problem: cannot access group directory
> 
> Hello,
> 
> I have a user `bbt_landing` who belongs to the group `builder`, and
> a directory name `test` which has the permission 750. The directory
> is located on a MFS volume. The problem is that user `bbt_landing`
> cannot access to the directory.
> 
> More details can be found at
> 	http://metakyanh.sarovar.org/moosefs.report/report3.html
> 
> Any idea?
> 


-- 
Anh Ky Huynh

Re: [Moosefs-users] problem: cannot access group directory

[Michał B. <mic...@ge...>]

Please check this FAQ entry:
http://www.moosefs.org/moosefs-faq.html#supplementary_groups


Regards
Michał 

-----Original Message-----
From: Anh K. Huynh [mailto:ky...@vi...] 
Sent: Wednesday, September 22, 2010 10:52 AM
To: moo...@li...
Subject: [Moosefs-users] problem: cannot access group directory

Hello,

I have a user `bbt_landing` who belongs to the group `builder`, and a
directory name `test` which has the permission 750. The directory is located
on a MFS volume. The problem is that user `bbt_landing` cannot access to the
directory.

More details can be found at
	http://metakyanh.sarovar.org/moosefs.report/report3.html

Any idea?

-- 
Anh Ky Huynh

----------------------------------------------------------------------------
--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
moosefs-users mailing list
moo...@li...
https://lists.sourceforge.net/lists/listinfo/moosefs-users

[Moosefs-users] problem: cannot access group directory

[Anh K. H. <ky...@vi...>]

Hello,

I have a user `bbt_landing` who belongs to the group `builder`, and a directory name `test` which has the permission 750. The directory is located on a MFS volume. The problem is that user `bbt_landing` cannot access to the directory.

More details can be found at
	http://metakyanh.sarovar.org/moosefs.report/report3.html

Any idea?

-- 
Anh Ky Huynh

Re: [Moosefs-users] duplicate entries in mounts information

[Anh K. H. <ky...@vi...>]

On Tue, 21 Sep 2010 16:37:59 +0200
Michał Borychowski <mic...@ge...> wrote:

> Check if you don't have any old "mfsmount" process?

Thanks so much.

Actually there's an old "mfsmount" on the system. That process was killed with -9 flag, and now `mfs.cgi` reports correctly.

:)

> 
> -----Original Message-----
> From: Anh K. Huynh [mailto:ky...@vi...] 
> Sent: Monday, September 20, 2010 4:25 AM
> To: MFS
> Subject: [Moosefs-users] duplicate entries in mounts information
> 
> Hello,
> 
> In my system, `mfs.cgi` reports two mount entries which are the
> same:
> 
> 13 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...
> 42 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...
> 
> I tried to restart mfsmaster, mfschunk, and then remount /0.MFS, but
> `mfs.cgi` still reported a same result. This is really strange, as
> other mount points are reported correctly.
> 
> Any idea?
> 


-- 
Anh Ky Huynh

Re: [Moosefs-users] duplicate entries in mounts information

[Michał B. <mic...@ge...>]

Check if you don't have any old "mfsmount" process?


Regards
Michał 

-----Original Message-----
From: Anh K. Huynh [mailto:ky...@vi...] 
Sent: Monday, September 20, 2010 4:25 AM
To: MFS
Subject: [Moosefs-users] duplicate entries in mounts information

Hello,

In my system, `mfs.cgi` reports two mount entries which are the same:

13 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...
42 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...

I tried to restart mfsmaster, mfschunk, and then remount /0.MFS, but
`mfs.cgi` still reported a same result. This is really strange, as other
mount points are reported correctly.

Any idea?

-- 
Anh Ky Huynh

----------------------------------------------------------------------------
--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
moosefs-users mailing list
moo...@li...
https://lists.sourceforge.net/lists/listinfo/moosefs-users

Re: [Moosefs-users] changelog on Metalogger not flush to disk instantly

[Michał B. <mic...@ge...>]

Thank you for your remark. Now in the development branch we’ve made automatic data flushing every second. Please also have a look at this thread:

http://sourceforge.net/mailarchive/forum.php?thread_name=00ba01cb5336%24e5765cf0%24b06316d0%24%40borychowski%40gemius.pl <http://sourceforge.net/mailarchive/forum.php?thread_name=00ba01cb5336%24e5765cf0%24b06316d0%24%40borychowski%40gemius.pl&forum_name=moosefs-users> &forum_name=moosefs-users

 

Kind regards

Michał 

 

From: 郑海洪 [mailto:zh...@uc...] 
Sent: Tuesday, September 21, 2010 3:58 AM
To: moosefs-users
Subject: [Moosefs-users] changelog on Metalogger not flush to disk instantly

 

Dear Michał,

 

I found the changelog on Metalogger machine not flush to disk instantly. I dig into the code, and found there miss a fflush in mfs-1.6.17/mfsmetalogger/masterconn.c:185.

 

142 void masterconn_metachanges_log(masterconn *eptr,const uint8_t *data,uint32_t length) {

143     char logname1[100],logname2[100];

.....

182     data++;

183     version = get64bit(&data);

184     if (eptr->logfd) {

185         fprintf(eptr->logfd,"%"PRIu64": %s\n",version,data);

              fflush(eptr->logfd);  //missed

186     } else {

187         syslog(LOG_NOTICE,"lost MFS change %"PRIu64": %s",version,data);

188     } 

189 }   

 

I think this maybe a tiny bug, because the changelog is so important to the metalogger, we need to flush it instantly.

 

Best regards!

 

Haihong Zheng from China

 

[Moosefs-users] changelog on Metalogger not flush to disk instantly

[郑. <zh...@uc...>]

Dear Michał,

I found the changelog on Metalogger machine not flush to disk instantly. I dig into the code, and found there miss a fflush in mfs-1.6.17/mfsmetalogger/masterconn.c:185.

142 void masterconn_metachanges_log(masterconn *eptr,const uint8_t *data,uint32_t length) {
143     char logname1[100],logname2[100];
.....
182     data++;
183     version = get64bit(&data);
184     if (eptr->logfd) {
185         fprintf(eptr->logfd,"%"PRIu64": %s\n",version,data);
              fflush(eptr->logfd);  //missed
186     } else {
187         syslog(LOG_NOTICE,"lost MFS change %"PRIu64": %s",version,data);
188     } 
189 }   

I think this maybe a tiny bug, because the changelog is so important to the metalogger, we need to flush it instantly.

Best regards!

Haihong Zheng from China

Re: [Moosefs-users] ask about storage backend for chunk servers

[Anh K. H. <ky...@vi...>]

On Mon, 20 Sep 2010 21:11:43 +0700
Cuong Hoang Bui <bhc...@gm...> wrote:

> I'm wondering about the storage backend for chunk servers. I have 2
> options 1. Using RAID for underly storage backend for chunk servers.
> 2. Not use RAID, Just separate disks attached directly to OS.
> 
> With RAID (maybe RAID 6) option, the solution is very good, but
> take money for RAID card.
> 
> With no RAID, I wonder about tolerance (disk failure). If I have 8
> disks attached directly to OS, for example I mount
> as /mnt/sdb, /mnt/sdc,.., .mnt/sdh. If one of them failed, what
> happens to data, also operations that's performing on this disk.
> Please give me some advice.

IMHO things depend on your "goal". If your goal is 2 or more, RAID isn't necessary (here the term "distributed"). As recommended from MFS docs, you should use RAID for mfsmaster/logger.

-- 
Anh Ky Huynh

[Moosefs-users] ask about storage backend for chunk servers

[Cuong H. B. <bhc...@gm...>]

Hi,

I'm wondering about the storage backend for chunk servers. I have 2 options
1. Using RAID for underly storage backend for chunk servers.
2. Not use RAID, Just separate disks attached directly to OS.

With RAID (maybe RAID 6) option, the solution is very good, but take 
money for RAID card.

With no RAID, I wonder about tolerance (disk failure). If I have 8 disks 
attached directly to OS, for example I mount as /mnt/sdb, /mnt/sdc,.., 
.mnt/sdh. If one of them failed, what happens to data, also operations 
that's performing on this disk. Please give me some advice.

-- 
**********************
Regards,
Cuong Hoang Bui
ct...@ct...
bhc...@gm...
**********************

[Moosefs-users] permission to access mfsmaster from mfslogger

[Anh K. H. <ky...@vi...>]

Hi,

What's the correct configuration in `mfsexports.cfg`, so that mfslogger can be able to access `mfsmaster` server? The "meta" option in `mfsexports.cfg` allows client to use meta information (thrash information), but it doesn't mention any stuff related to logger. 

Thank you for your replies,

-- 
Anh Ky Huynh

[Moosefs-users] duplicate entries in mounts information

[Anh K. H. <ky...@vi...>]

Hello,

In my system, `mfs.cgi` reports two mount entries which are the same:

13 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...
42 db1.internal a.b.c.d /0.MFS 1.6.17 /db1 rw...

I tried to restart mfsmaster, mfschunk, and then remount /0.MFS, but `mfs.cgi` still reported a same result. This is really strange, as other mount points are reported correctly.

Any idea?

-- 
Anh Ky Huynh

[Moosefs-users] Hypertable on MooseFS

[Kristofer P. <kri...@cy...>]

Just curious if anyone is using Hypertable on MFS, or has tested it?

If so, I would like to ask some questions about the implementation of it.  I realize that Hypertable questions should be asked on the Hypertable list, but since this is more specific to MooseFS, I thought there may be some people on this list that may have experience with it.

Thanks,
Kris

Re: [Moosefs-users] What companies use MooseFS

[Anh K. H. <ky...@vi...>]

On Wed, 15 Sep 2010 13:50:47 +0200
Ioannis Aslanidis <ias...@fl...> wrote:

> I am wondering if there is any list of companies that use MooseFS
> for their infrastructure. I need to know how widely spread MooseFS
> is before being able to implement it. Can anyone provide info on
> that?

We're Skunkworks at Vietnam (skunkworks.vn); we have a small MFS setup with 1 master, 1 metalogger and 2 chunk servers (these chunk servers are also the clients.) The primary purpose is to replicate and share data between machines (which are provided by Amazon, cf. http://aws.amazon.com/ec2). We found that NFS was so complex, so we choose MFS ;) Our current storage is only 186 GB, but it will be expanded soon.

As our servers are running in a cloud/virtual environments, the speed varies and depends on global load: for writing, it varies from 3MiB/s -> 154 MiB/s (the maximum value that I experienced.)

Have fun,

-- 
Anh Ky Huynh

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Fabien G. <fab...@gm...>]

On Thu, Sep 16, 2010 at 10:00 AM, Laurent Wandrebeck <lw...@hy...> wrote:

> > Nice you're talking about the trash bin :) : I was asking myself how to
> get
> > access to its content. The documentation only says " Removed files may be
> > accessed through a separately mounted MFSMETA file system". What does it
> > mean ?
> http://www.moosefs.org/reference-guide.html#using-moosefs
> « By starting the mfsmount process with the -m (or -o mfsmeta) option
> one can mount the auxiliary file system MFSMETA (which may be useful to
> restore a file accidentally deleted from the MooseFS volume or to free
> some space by removing a file before elapsing the quarantine time), for
> example:
>
> mfsmount -m /mnt/mfsmeta »
> add -H mfsmaster_host and it works (just tested:)


Yes, that's exactly what I was looking for : "-o mfsmeta". It works, I now
can reach the reserved+trash directories, thank you !

Fabien

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Michał B. <mic...@ge...>]

Yes, this is the right part of documentation. At
http://www.moosefs.org/reference-guide.html find: “Removed files may be
accessed through a separately mounted MFSMETA file system” and keep on
reading.  Below there is: “Moving this file to the trash/undel
subdirectory causes restoring of the original file in a proper MooseFS file
system”

 

 

Regards

Michał 

 

From: Fabien Germain [mailto:fab...@gm...] 
Sent: Thursday, September 16, 2010 9:52 AM
To: Laurent Wandrebeck
Cc: moo...@li...; 夏亮
Subject: Re: [Moosefs-users] rm the size of chunkserver is not changed

 

Hello,

On Thu, Sep 16, 2010 at 8:35 AM, Laurent Wandrebeck <lw...@hy...> wrote:

>          When I rm a big file of 80G. I found that the disk size of
> chunkserver didn’t changed
>
> Why?

Because of the trash :)
See http://www.moosefs.org/reference-guide.html#using-moosefs , basic
operations part.


Nice you're talking about the trash bin :) : I was asking myself how to get
access to its content. The documentation only says " Removed files may be
accessed through a separately mounted MFSMETA file system". What does it
mean ?

Fabien

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Laurent W. <lw...@hy...>]

On Thu, 16 Sep 2010 09:51:40 +0200
Fabien Germain <fab...@gm...> wrote:

> Hello,
> 
> On Thu, Sep 16, 2010 at 8:35 AM, Laurent Wandrebeck <lw...@hy...> wrote:
> 
> > >          When I rm a big file of 80G. I found that the disk size of
> > > chunkserver didn’t changed
> > >
> > > Why?
> > Because of the trash :)
> > See http://www.moosefs.org/reference-guide.html#using-moosefs , basic
> > operations part.
> >
> 
> Nice you're talking about the trash bin :) : I was asking myself how to get
> access to its content. The documentation only says " Removed files may be
> accessed through a separately mounted MFSMETA file system". What does it
> mean ?
http://www.moosefs.org/reference-guide.html#using-moosefs
« By starting the mfsmount process with the -m (or -o mfsmeta) option
one can mount the auxiliary file system MFSMETA (which may be useful to
restore a file accidentally deleted from the MooseFS volume or to free
some space by removing a file before elapsing the quarantine time), for
example:

mfsmount -m /mnt/mfsmeta »
add -H mfsmaster_host and it works (just tested:)
Hope it helps,
-- 
Laurent Wandrebeck
HYGEOS, Earth Observation Department / Observation de la Terre
Euratechnologies
165 Avenue de Bretagne
59000 Lille, France
tel: +33 3 20 08 24 98
http://www.hygeos.com
GPG fingerprint/Empreinte GPG: F5CA 37A4 6D03 A90C 7A1D  2A62 54E6 EF2C
D17C F64C

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Fabien G. <fab...@gm...>]

Hello,

On Thu, Sep 16, 2010 at 8:35 AM, Laurent Wandrebeck <lw...@hy...> wrote:

> >          When I rm a big file of 80G. I found that the disk size of
> > chunkserver didn’t changed
> >
> > Why?
> Because of the trash :)
> See http://www.moosefs.org/reference-guide.html#using-moosefs , basic
> operations part.
>

Nice you're talking about the trash bin :) : I was asking myself how to get
access to its content. The documentation only says " Removed files may be
accessed through a separately mounted MFSMETA file system". What does it
mean ?

Fabien

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Michał B. <mic...@ge...>]

Please have a look here:

http://www.moosefs.org/moosefs-faq.html#delete

 

 

Kind regards

Michał Borychowski 

MooseFS Support Manager

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Gemius S.A.

ul. Wołoska 7, 02-672 Warszawa

Budynek MARS, klatka D

Tel.: +4822 874-41-00

Fax : +4822 874-41-01

 

From: 夏亮 [mailto:xia...@zh...] 
Sent: Thursday, September 16, 2010 5:48 AM
To: moo...@li...
Subject: [Moosefs-users] rm the size of chunkserver is not changed

 

Hi :

         When I rm a big file of 80G. I found that the disk size of
chunkserver didn’t changed 

Why?

Re: [Moosefs-users] What companies use MooseFS

[Michał B. <mic...@ge...>]

And there is Gemius (http://www.gemius.com) with four deployments, the biggest has almost 30 million files distributed over 70 chunk servers having a total space of 570TiB. Chunkserver machines at the same time are used to make other calculations. You can download some screenshots from our CGI monitor here:
http://www.moosefs.org/tl_files/monitor_screens_200912.tar

Another Polish companies which uses MooseFS for data storage are Redefine (http://www.redefine.pl/) and AdOcean (http://www.adocean-global.com/).


Kind regards
Michał Borychowski 
MooseFS Support Manager
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Gemius S.A.
ul. Wołoska 7, 02-672 Warszawa
Budynek MARS, klatka D
Tel.: +4822 874-41-00
Fax : +4822 874-41-01



-----Original Message-----
From: jose maria [mailto:let...@us...] 
Sent: Thursday, September 16, 2010 2:14 AM
To: moo...@li...
Subject: Re: [Moosefs-users] What companies use MooseFS


> 
>> Hello,
>>
>> I am wondering if there is any list of companies that use MooseFS for
>> their infrastructure. I need to know how widely spread MooseFS is
>> before being able to implement it. Can anyone provide info on that?
>>
>> Thank you.
>>

* www.seycob.es , www.copiadeseguridad.cl , remote backup company's

* Four mfs clusters , mfsclients on root vsftpd virtual-servers, example 
on spain, primary cluster http://control.seycob.es:9425 , rsync to
secondary cluster, http://control.seycob.es:9426
are on the same subnet, but in different locations, connected trough 
fiber optics.

* Absolute stability in the processes 24/7 continuous read/write for ten 
months in operation, seven discs damaged, ext3 filesystem, problems on 
ext4 and btrfs. I'm trying again in ext4 ....

* Six sata disks +- per chunkserver, comodity hardware HP/DELL
+- 280 € + discs per server, except mfsmaster's, 2 Gigabit Ethernet with 
Link Agregation in mode 5. Moosefs has fulfilled our expectations 
provided stability, performance, scalability, consistency, fault 
tolerance, content costs, communication with applications, management 
simple..

* At first I was evaluating the deployment of cluster Cassandra, but 
stability and access for applications was painful, I am not a fan of 
java ...., but is now a viable option for file-based nosql/bigtable 
cluster. Or glusterfs for FTND raid1 filesystem.

* Depending on the goal, fortunately there are several open source 
options. MooseFS is the most obvious chunk-based cluster.

* Pending features in the future, secondary mfsmaster, define Rac's on 
mfsmaster/mfschunserver, goal per rac, mfstools for massive manage 
errors in chunks, inodes, files, reserverd files, mfssnapshot is 
unusable for massive or continuous backup purposes.

* Regards.


------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
moosefs-users mailing list
moo...@li...
https://lists.sourceforge.net/lists/listinfo/moosefs-users

Re: [Moosefs-users] rm the size of chunkserver is not changed

[Laurent W. <lw...@hy...>]

On Thu, 16 Sep 2010 11:47:40 +0800
夏亮 <xia...@zh...> wrote:

> Hi :
> 
>          When I rm a big file of 80G. I found that the disk size of
> chunkserver didn’t changed 
> 
> Why?
Because of the trash :)
See http://www.moosefs.org/reference-guide.html#using-moosefs , basic
operations part.
Regards,
-- 
Laurent Wandrebeck
HYGEOS, Earth Observation Department / Observation de la Terre
Euratechnologies
165 Avenue de Bretagne
59000 Lille, France
tel: +33 3 20 08 24 98
http://www.hygeos.com
GPG fingerprint/Empreinte GPG: F5CA 37A4 6D03 A90C 7A1D  2A62 54E6 EF2C
D17C F64C

[Moosefs-users] rm the size of chunkserver is not changed

[夏亮 <xia...@zh...>]

Hi :

         When I rm a big file of 80G. I found that the disk size of
chunkserver didn’t changed 

Why?

Re: [Moosefs-users] What companies use MooseFS

[jose m. <let...@us...>]

> 
>> Hello,
>>
>> I am wondering if there is any list of companies that use MooseFS for
>> their infrastructure. I need to know how widely spread MooseFS is
>> before being able to implement it. Can anyone provide info on that?
>>
>> Thank you.
>>

* www.seycob.es , www.copiadeseguridad.cl , remote backup company's

* Four mfs clusters , mfsclients on root vsftpd virtual-servers, example 
on spain, primary cluster http://control.seycob.es:9425 , rsync to
secondary cluster, http://control.seycob.es:9426
are on the same subnet, but in different locations, connected trough 
fiber optics.

* Absolute stability in the processes 24/7 continuous read/write for ten 
months in operation, seven discs damaged, ext3 filesystem, problems on 
ext4 and btrfs. I'm trying again in ext4 ....

* Six sata disks +- per chunkserver, comodity hardware HP/DELL
+- 280 € + discs per server, except mfsmaster's, 2 Gigabit Ethernet with 
Link Agregation in mode 5. Moosefs has fulfilled our expectations 
provided stability, performance, scalability, consistency, fault 
tolerance, content costs, communication with applications, management 
simple..

* At first I was evaluating the deployment of cluster Cassandra, but 
stability and access for applications was painful, I am not a fan of 
java ...., but is now a viable option for file-based nosql/bigtable 
cluster. Or glusterfs for FTND raid1 filesystem.

* Depending on the goal, fortunately there are several open source 
options. MooseFS is the most obvious chunk-based cluster.

* Pending features in the future, secondary mfsmaster, define Rac's on 
mfsmaster/mfschunserver, goal per rac, mfstools for massive manage 
errors in chunks, inodes, files, reserverd files, mfssnapshot is 
unusable for massive or continuous backup purposes.

* Regards.

Re: [Moosefs-users] What companies use MooseFS

[Laurent W. <lw...@hy...>]

On Wed, 15 Sep 2010 14:22:37 +0200
Bán Miklós <ba...@vo...> wrote:

> Hi, we recently started to use it in University of Debrecen, Hungary on
> 7 machines ~ 3Tb data.
Hi,
I'm right now deploying it after a testing phase. moving 70TB takes a
bit of time ;)
-- 
Laurent Wandrebeck
HYGEOS, Earth Observation Department / Observation de la Terre
Euratechnologies
165 Avenue de Bretagne
59000 Lille, France
tel: +33 3 20 08 24 98
http://www.hygeos.com
GPG fingerprint/Empreinte GPG: F5CA 37A4 6D03 A90C 7A1D  2A62 54E6 EF2C
D17C F64C