moosefs-users — Mailing list for MooseFS Users

[MooseFS-Users] "An unexpected network error occurred"

[Neddy, N. N. <na...@nd...>]

Hi,

I got this error message when copying data from a Novell network drive
to SMB drive (that's a MooseFS client).

The topology:

PC----(LAN)-----SAMBA-----MooseFS cluster
          |
       Novell Netware

The PC was mapping a few network drives both from Novell and Samba
servers, and it's a Windows 7 SP1 x64. So, I opened a Novell drive and
chose all and "Send To" Samba drive, then a moment later this message
appears "0x8007003b, An unexpected network error occurred.". Google
led to MS page https://support.microsoft.com/en-us/kb/2649905 but
applied hotfix didn't resolve the error.

It doesn't look like MooseFS's fault, but I appreciate your comments
on this case.

Thanks,
~Nedd

Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Eduardo K. <edu...@do...>]

 Krzysztof Kielak, 

thank you!!

Best Regards,

Eduardo Kellenberger


El Lunes 06 abril 2015 19:33:17 Krzysztof Kielak escribió:
> Dear Eduardo,
> 
> We'll look at this issue definitely tomorrow. It should be possible to
> change those parameters in MooseFS version 2.0.
> 
> Best Regards,
> Krzysztof Kielak
> Director of Operations and Customer Support
> Mobile: +48 601 476 440
> 
> -----Original Message-----
> From: Eduardo Kellenberger [mailto:edu...@do...] 
> Sent: Monday, April 6, 2015 2:02 PM
> To: Aleksander Wieliczko
> Cc: moo...@li...
> Subject: Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT
> on moosefs ce
> 
> Hi, thanks for your answer.
> 
> The number of erasure is variable, I neesito is that from 8am to 11pm files
> are not deleted, or the least amount is removed to prevent overloading of
> chunckserver and between 11pm and 8am deletion rise.
> 
> This I could control with version 1.6 but with the 2 EC no.
> 
> 
> best regard, 
> 
> Eduardo Kellenberger
> 
> 
> 
> 
> 
> El Jueves 02 abril 2015 10:33:34 Aleksander Wieliczko escribió:
> > Hello
> > First of all I would like to explain what these parameters mean and 
> > how does it work:
> > 
> > CHUNKS_SOFT_DEL_LIMIT - Soft maximum number of chunks being 
> > simultaneously deleted on one chunkserver (default is 10) 
> > CHUNKS_HARD_DEL_LIMIT - Hard maximum number of chunks being 
> > simultaneously deleted on one chunkserver (default is 25)
> > 
> > MooseFS mainly uses the soft delete limit. If the system is able to 
> > remove all "to delete" chunks in one loop run, deletion limit will not 
> > be changed.
> > But if the number of chunks to delete increase after a loop run, 
> > deletion limit will be increased by factor of 1.5 but never to a value 
> > higher than CHUNKS_HARD_DEL_LIMIT.
> > 
> > This mechanism was implemented to prevent MooseFS cluster from 
> > freezing because of too many delete operations.
> > So if you have small number of delete operations in your system hard 
> > limit will never be reached.
> > 
> > How many files are you deleting in your system?
> 
> --
> -------------------------------------------------
> Eduardo Kellenberger
> Departamento de Infraestructura Tecnológica DonWeb "La actitud es todo"
> Donweb.com
> 
> Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son
> confidenciales, de uso exclusivo para el destinatario del mismo. La
> divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda
> prohibida.
> DonWeb.com no se hace responsable del mensaje por la falsificación y/o
> alteración del mismo.
> De no ser Ud el destinatario del mismo y lo ha recibido por error, por
> favor, notifique al remitente y elimínelo de su sistema.
> 
> Confidentiality Note: This message and any attachments (the message) are
> confidential and intended solely for the addressees. Any unauthorised use or
> dissemination is prohibited by DonWeb.com.
> DonWeb.com shall not be liable  for the message if altered or falsified.
> If you are not the intended addressee of this message, please cancel it
> immediately and inform the sender
> 
> Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem
> conter dados confidenciais ou privilegiados.
> Se você os recebeu por engano ou não é um dos destinatários aos quais ela
> foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou
> copias realizadas, imediatamente.
> É proibida a retenção, distribuição, divulgação ou utilização de quaisquer
> informações aqui contidas.
> 
> Por favor, informenos sobre o recebimento indevido desta mensagem,
> retornando-a para o autor.
> 
> 
> ----------------------------------------------------------------------------
> --
> BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your
> own process in accordance with the BPMN 2 standard Learn Process modeling
> best practices with Bonita BPM through live exercises
> http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
> source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
> 
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
> 

-- 
-------------------------------------------------
Eduardo Kellenberger
Departamento de Infraestructura Tecnológica
DonWeb
"La actitud es todo"
Donweb.com

Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son confidenciales, de uso exclusivo para el destinatario del mismo. La divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda prohibida.
DonWeb.com no se hace responsable del mensaje por la falsificación y/o alteración del mismo.
De no ser Ud el destinatario del mismo y lo ha recibido por error, por favor, notifique al remitente y elimínelo de su sistema.

Confidentiality Note: This message and any attachments (the message) are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited by DonWeb.com.
DonWeb.com shall not be liable  for the message if altered or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender

Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem conter dados confidenciais ou privilegiados.
Se você os recebeu por engano ou não é um dos destinatários aos quais ela foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou copias realizadas, imediatamente.
É proibida a retenção, distribuição, divulgação ou utilização de quaisquer informações aqui contidas.

Por favor, informenos sobre o recebimento indevido desta mensagem, retornando-a para o autor.

Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Krzysztof K. <krz...@mo...>]

Dear Eduardo,

We'll look at this issue definitely tomorrow. It should be possible to
change those parameters in MooseFS version 2.0.

Best Regards,
Krzysztof Kielak
Director of Operations and Customer Support
Mobile: +48 601 476 440

-----Original Message-----
From: Eduardo Kellenberger [mailto:edu...@do...] 
Sent: Monday, April 6, 2015 2:02 PM
To: Aleksander Wieliczko
Cc: moo...@li...
Subject: Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT
on moosefs ce

Hi, thanks for your answer.

The number of erasure is variable, I neesito is that from 8am to 11pm files
are not deleted, or the least amount is removed to prevent overloading of
chunckserver and between 11pm and 8am deletion rise.

This I could control with version 1.6 but with the 2 EC no.


best regard, 

Eduardo Kellenberger





El Jueves 02 abril 2015 10:33:34 Aleksander Wieliczko escribió:
> Hello
> First of all I would like to explain what these parameters mean and 
> how does it work:
> 
> CHUNKS_SOFT_DEL_LIMIT - Soft maximum number of chunks being 
> simultaneously deleted on one chunkserver (default is 10) 
> CHUNKS_HARD_DEL_LIMIT - Hard maximum number of chunks being 
> simultaneously deleted on one chunkserver (default is 25)
> 
> MooseFS mainly uses the soft delete limit. If the system is able to 
> remove all "to delete" chunks in one loop run, deletion limit will not 
> be changed.
> But if the number of chunks to delete increase after a loop run, 
> deletion limit will be increased by factor of 1.5 but never to a value 
> higher than CHUNKS_HARD_DEL_LIMIT.
> 
> This mechanism was implemented to prevent MooseFS cluster from 
> freezing because of too many delete operations.
> So if you have small number of delete operations in your system hard 
> limit will never be reached.
> 
> How many files are you deleting in your system?

--
-------------------------------------------------
Eduardo Kellenberger
Departamento de Infraestructura Tecnológica DonWeb "La actitud es todo"
Donweb.com

Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son
confidenciales, de uso exclusivo para el destinatario del mismo. La
divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda
prohibida.
DonWeb.com no se hace responsable del mensaje por la falsificación y/o
alteración del mismo.
De no ser Ud el destinatario del mismo y lo ha recibido por error, por
favor, notifique al remitente y elimínelo de su sistema.

Confidentiality Note: This message and any attachments (the message) are
confidential and intended solely for the addressees. Any unauthorised use or
dissemination is prohibited by DonWeb.com.
DonWeb.com shall not be liable  for the message if altered or falsified.
If you are not the intended addressee of this message, please cancel it
immediately and inform the sender

Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem
conter dados confidenciais ou privilegiados.
Se você os recebeu por engano ou não é um dos destinatários aos quais ela
foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou
copias realizadas, imediatamente.
É proibida a retenção, distribuição, divulgação ou utilização de quaisquer
informações aqui contidas.

Por favor, informenos sobre o recebimento indevido desta mensagem,
retornando-a para o autor.


----------------------------------------------------------------------------
--
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your
own process in accordance with the BPMN 2 standard Learn Process modeling
best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF

_________________________________________
moosefs-users mailing list
moo...@li...
https://lists.sourceforge.net/lists/listinfo/moosefs-users

Re: [MooseFS-Users] moosefs-users Digest, Vol 64, Issue 3

[Steve W. <st...@pu...>]

Oh, okay... sorry I missed that you were already doing that.

Steve

On 04/06/2015 09:20 AM, Eduardo Kellenberger wrote:
> Steve,
>
> It's exactly what I do, but I see no difference between the two config.-
>
> As indicated earlier, this same work correctly in version 1.6
>
> Best regard
>
> Eduardo
>
>
>
> El Lunes 06 abril 2015 12:21:34 moo...@li... escribió:
>> ------------------------------
>>
>> Message: 5
>> Date: Mon, 06 Apr 2015 08:21:26 -0400
>> From: Steve Wilson <st...@pu...>
>> Subject: Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and
>>          CHUNKS_HARD_DEL_LIMIT on moosefs ce
>> To: moo...@li...
>> Message-ID: <552...@pu...>
>> Content-Type: text/plain; charset=windows-1252; format=flowed
>>
>> Perhaps you could have a cron job that would run every day at 8 a.m. and
>> another one at 11 p.m.  The 8 a.m. cron job would create a symbolic link
>> from /etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-day.cfg and then send
>> a signal to the mfsmaster to reload its configuration.  And similarly,
>> the 11 p.m. cron job would create a symbolic link from
>> /etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-night.cfg and send a reload
>> signal to mfsmaster. The two different config files would, of course,
>> have the two different settings that you want, one for the daytime hours
>> and another for the nighttime hours.
>>
>> Steve

Re: [MooseFS-Users] moosefs-users Digest, Vol 64, Issue 3

[Eduardo K. <edu...@do...>]

Steve, 

It's exactly what I do, but I see no difference between the two config.-

As indicated earlier, this same work correctly in version 1.6

Best regard

Eduardo



El Lunes 06 abril 2015 12:21:34 moo...@li... escribió:
> 
> ------------------------------
> 
> Message: 5
> Date: Mon, 06 Apr 2015 08:21:26 -0400
> From: Steve Wilson <st...@pu...>
> Subject: Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and
>         CHUNKS_HARD_DEL_LIMIT on moosefs ce
> To: moo...@li...
> Message-ID: <552...@pu...>
> Content-Type: text/plain; charset=windows-1252; format=flowed
> 
> Perhaps you could have a cron job that would run every day at 8 a.m. and 
> another one at 11 p.m.  The 8 a.m. cron job would create a symbolic link 
> from /etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-day.cfg and then send 
> a signal to the mfsmaster to reload its configuration.  And similarly, 
> the 11 p.m. cron job would create a symbolic link from 
> /etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-night.cfg and send a reload 
> signal to mfsmaster. The two different config files would, of course, 
> have the two different settings that you want, one for the daytime hours 
> and another for the nighttime hours.
> 
> Steve

-- 
-------------------------------------------------
Eduardo Kellenberger
Departamento de Infraestructura Tecnológica
DonWeb
"La actitud es todo"
Donweb.com

Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son confidenciales, de uso exclusivo para el destinatario del mismo. La divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda prohibida.
DonWeb.com no se hace responsable del mensaje por la falsificación y/o alteración del mismo.
De no ser Ud el destinatario del mismo y lo ha recibido por error, por favor, notifique al remitente y elimínelo de su sistema.

Confidentiality Note: This message and any attachments (the message) are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited by DonWeb.com.
DonWeb.com shall not be liable  for the message if altered or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender

Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem conter dados confidenciais ou privilegiados.
Se você os recebeu por engano ou não é um dos destinatários aos quais ela foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou copias realizadas, imediatamente.
É proibida a retenção, distribuição, divulgação ou utilização de quaisquer informações aqui contidas.

Por favor, informenos sobre o recebimento indevido desta mensagem, retornando-a para o autor.

Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Steve W. <st...@pu...>]

Perhaps you could have a cron job that would run every day at 8 a.m. and 
another one at 11 p.m.  The 8 a.m. cron job would create a symbolic link 
from /etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-day.cfg and then send 
a signal to the mfsmaster to reload its configuration.  And similarly, 
the 11 p.m. cron job would create a symbolic link from 
/etc/mfs/mfsmaster.cfg to /etc/mfs/mfsmaster-night.cfg and send a reload 
signal to mfsmaster. The two different config files would, of course, 
have the two different settings that you want, one for the daytime hours 
and another for the nighttime hours.

Steve

On 04/06/2015 08:01 AM, Eduardo Kellenberger wrote:
> Hi, thanks for your answer.
>
> The number of erasure is variable, I neesito is that from 8am to 11pm files are not deleted, or the least amount is removed to prevent overloading of chunckserver and between 11pm and 8am deletion rise.
>
> This I could control with version 1.6 but with the 2 EC no.
>
>
> best regard,
>
> Eduardo Kellenberger
>
>
>
>
>
> El Jueves 02 abril 2015 10:33:34 Aleksander Wieliczko escribió:
>> Hello
>> First of all I would like to explain what these parameters mean and how
>> does it work:
>>
>> CHUNKS_SOFT_DEL_LIMIT - Soft maximum number of chunks being
>> simultaneously deleted on one chunkserver (default is 10)
>> CHUNKS_HARD_DEL_LIMIT - Hard maximum number of chunks being
>> simultaneously deleted on one chunkserver (default is 25)
>>
>> MooseFS mainly uses the soft delete limit. If the system is able to
>> remove all "to delete" chunks in one loop run, deletion limit will not
>> be changed.
>> But if the number of chunks to delete increase after a loop run,
>> deletion limit will be increased by factor of 1.5 but never to a value
>> higher than CHUNKS_HARD_DEL_LIMIT.
>>
>> This mechanism was implemented to prevent MooseFS cluster from freezing
>> because of too many delete operations.
>> So if you have small number of delete operations in your system hard
>> limit will never be reached.
>>
>> How many files are you deleting in your system?

Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Eduardo K. <edu...@do...>]

Hi, thanks for your answer.

The number of erasure is variable, I neesito is that from 8am to 11pm files are not deleted, or the least amount is removed to prevent overloading of chunckserver and between 11pm and 8am deletion rise.

This I could control with version 1.6 but with the 2 EC no.


best regard, 

Eduardo Kellenberger





El Jueves 02 abril 2015 10:33:34 Aleksander Wieliczko escribió:
> Hello
> First of all I would like to explain what these parameters mean and how
> does it work:
> 
> CHUNKS_SOFT_DEL_LIMIT - Soft maximum number of chunks being
> simultaneously deleted on one chunkserver (default is 10)
> CHUNKS_HARD_DEL_LIMIT - Hard maximum number of chunks being
> simultaneously deleted on one chunkserver (default is 25)
> 
> MooseFS mainly uses the soft delete limit. If the system is able to
> remove all "to delete" chunks in one loop run, deletion limit will not
> be changed.
> But if the number of chunks to delete increase after a loop run,
> deletion limit will be increased by factor of 1.5 but never to a value
> higher than CHUNKS_HARD_DEL_LIMIT.
> 
> This mechanism was implemented to prevent MooseFS cluster from freezing
> because of too many delete operations.
> So if you have small number of delete operations in your system hard
> limit will never be reached.
> 
> How many files are you deleting in your system?

-- 
-------------------------------------------------
Eduardo Kellenberger
Departamento de Infraestructura Tecnológica
DonWeb
"La actitud es todo"
Donweb.com

Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son confidenciales, de uso exclusivo para el destinatario del mismo. La divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda prohibida.
DonWeb.com no se hace responsable del mensaje por la falsificación y/o alteración del mismo.
De no ser Ud el destinatario del mismo y lo ha recibido por error, por favor, notifique al remitente y elimínelo de su sistema.

Confidentiality Note: This message and any attachments (the message) are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited by DonWeb.com.
DonWeb.com shall not be liable  for the message if altered or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender

Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem conter dados confidenciais ou privilegiados.
Se você os recebeu por engano ou não é um dos destinatários aos quais ela foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou copias realizadas, imediatamente.
É proibida a retenção, distribuição, divulgação ou utilização de quaisquer informações aqui contidas.

Por favor, informenos sobre o recebimento indevido desta mensagem, retornando-a para o autor.

Re: [MooseFS-Users] Slow replication between internal disks of 1 chunkserver

[Aleksander W. <ale...@mo...>]

Hi.
Yes.

Of course every thing depend on your system characteristic but if you
have a lot of replications, other operations can be starved by internal
chunkserver rebalancing process.
Hold to the view that one chunkserver will be overloaded by internal
replication process, whole cluster will have less number of I/O operations.

We recommend using default settings for HDD_REBALANCE_UTILIZATION option.

Best regards
Aleksander Wieliczko
Technical Support Engineer
MooseFS.com <moosefs.com>

On 04/03/2015 10:39 AM, Neddy, NH. Nam wrote:
> I increased HDD_REBALANCE_UTILIZATION value and the performance is
> improved significantly. So, is there any drawback if this value is too
> high?
>
> Thanks,
>
> On Fri, Apr 3, 2015 at 3:16 PM, Aleksander Wieliczko
> <ale...@mo...> wrote:
>> Hello.
>>
>> Starting from version MooseFS 2.0.x there's a new mechanism for internal
>> replication within chunkserver that can be controlled using
>> HDD_REBALANCE_UTILIZATION parameter in mfschunkserver.cfg file.
>>
>> So you have two options:
>> 1. You can set it to minimal value of 1, to decrease the impact of internal
>> rebalancing on i/o performance of the chunkserver.
>> HDD_REBALANCE_UTILIZATION = 1
>>
>> 2. You can set it to higher number during off-peak hours so the rebalancing
>> between chunkservers disks will finish faster in some maintenance window.
>>
>> Best regards
>> Aleksander Wieliczko
>> Technical Support Engineer
>> MooseFS.com
>>
>>
>> On 04/03/2015 09:45 AM, Davies Liu wrote:
>>
>> You can have a script to copy some files into new disk and create a
>> symlink for copied files.
>>
>> Once the disks got balanced, you can stop the chunk server, REMOVE all
>> symlinks, then start chunk server.
>>
>> On Thu, Apr 2, 2015 at 11:30 PM, Neddy, NH. Nam <na...@nd...> wrote:
>>
>> Hi,
>>
>> One disk was added to a chunkserver (2.0.60 CE) recently, but the
>> performance seems to be not good compared to adding totally new
>> chunkserver.
>>
>> All the disks are the same capacity but different brands. The
>> previously new added chunkserver has performed at limit write speed of
>> SATA2 (avg 60 MB/s), but the current server with brand disk added that
>> only has been wrting avg. 25 MB/s. (as seen in server graphs)
>>
>> Does anybody have idea on this?
>> Thanks,
>> ~Neddy
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website,
>> sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub for
>> all
>> things parallel software development, from weekly thought leadership blogs
>> to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now. http://goparallel.sourceforge.net/
>>
>> _________________________________________
>> moosefs-users mailing list
>> moo...@li...
>> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>>
>>
>>

[MooseFS-Users] MooseFS Easter Edition !!

[Krzysztof K. <krz...@mo...>]

Dear MooseFS Community,

A new version of MooseFS is coming to you on Easter with some new exciting features and specially built packages for “MooseFS 3.0 Easter Edition for Raspberry Pi 2” running on Raspbian Wheezy :)

Among many updates and fixes, that are laboriously listed in source packages change log, here’s a quick view on most important changes and new features:

- Added new functionality for Storage Tiering (possibility to assign labels to chunkservers and use those labels in mfssetgoal expressions when defining policies for storing, keeping and archiving data),
- Performance improvement for small file random I/O due to the new semantics for fsync() system call, which now by default has the same behaviour as on any local Linux/FreeBSD filesystem,
- Support for global locks compatible with POSIX locks and flock() advisory lock mechanism when using fuse 2.9+.

Simple test of unpacking complete source tree for latest Linux kernel (more than 50k objects with cumulative size of more than 630 MB when unpacked) shows 100% speed improvement when going from MooseFS version 2.0.x to version 3.0.x.

Be sure to check the documentation on new storage tiering feature and installation instructions at http://moosefs.com.

Best Regards,
Krzysztof Kielak
Director of Operations and Customer Support
Mobile: +48 601 476 440

Re: [MooseFS-Users] Slow replication between internal disks of 1 chunkserver

[Neddy, N. N. <na...@nd...>]

I increased HDD_REBALANCE_UTILIZATION value and the performance is
improved significantly. So, is there any drawback if this value is too
high?

Thanks,

On Fri, Apr 3, 2015 at 3:16 PM, Aleksander Wieliczko
<ale...@mo...> wrote:
> Hello.
>
> Starting from version MooseFS 2.0.x there's a new mechanism for internal
> replication within chunkserver that can be controlled using
> HDD_REBALANCE_UTILIZATION parameter in mfschunkserver.cfg file.
>
> So you have two options:
> 1. You can set it to minimal value of 1, to decrease the impact of internal
> rebalancing on i/o performance of the chunkserver.
> HDD_REBALANCE_UTILIZATION = 1
>
> 2. You can set it to higher number during off-peak hours so the rebalancing
> between chunkservers disks will finish faster in some maintenance window.
>
> Best regards
> Aleksander Wieliczko
> Technical Support Engineer
> MooseFS.com
>
>
> On 04/03/2015 09:45 AM, Davies Liu wrote:
>
> You can have a script to copy some files into new disk and create a
> symlink for copied files.
>
> Once the disks got balanced, you can stop the chunk server, REMOVE all
> symlinks, then start chunk server.
>
> On Thu, Apr 2, 2015 at 11:30 PM, Neddy, NH. Nam <na...@nd...> wrote:
>
> Hi,
>
> One disk was added to a chunkserver (2.0.60 CE) recently, but the
> performance seems to be not good compared to adding totally new
> chunkserver.
>
> All the disks are the same capacity but different brands. The
> previously new added chunkserver has performed at limit write speed of
> SATA2 (avg 60 MB/s), but the current server with brand disk added that
> only has been wrting avg. 25 MB/s. (as seen in server graphs)
>
> Does anybody have idea on this?
> Thanks,
> ~Neddy
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>
>
>

Re: [MooseFS-Users] Slow replication between internal disks of 1 chunkserver

[Aleksander W. <ale...@mo...>]

Hello.

Starting from version MooseFS 2.0.x there's a new mechanism for internal
replication within chunkserver that can be controlled using
HDD_REBALANCE_UTILIZATION parameter in mfschunkserver.cfg file.
 
So you have two options:
1. You can set it to minimal value of 1, to decrease the impact of
internal rebalancing on i/o performance of the chunkserver.
HDD_REBALANCE_UTILIZATION = 1

2. You can set it to higher number during off-peak hours so the
rebalancing between chunkservers disks will finish faster in some
maintenance window.

Best regards
Aleksander Wieliczko
Technical Support Engineer
MooseFS.com <moosefs.com>


On 04/03/2015 09:45 AM, Davies Liu wrote:
> You can have a script to copy some files into new disk and create a
> symlink for copied files.
>
> Once the disks got balanced, you can stop the chunk server, REMOVE all
> symlinks, then start chunk server.
>
> On Thu, Apr 2, 2015 at 11:30 PM, Neddy, NH. Nam <na...@nd...> wrote:
>> Hi,
>>
>> One disk was added to a chunkserver (2.0.60 CE) recently, but the
>> performance seems to be not good compared to adding totally new
>> chunkserver.
>>
>> All the disks are the same capacity but different brands. The
>> previously new added chunkserver has performed at limit write speed of
>> SATA2 (avg 60 MB/s), but the current server with brand disk added that
>> only has been wrting avg. 25 MB/s. (as seen in server graphs)
>>
>> Does anybody have idea on this?
>> Thanks,
>> ~Neddy
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub for all
>> things parallel software development, from weekly thought leadership blogs to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now. http://goparallel.sourceforge.net/
>>
>> _________________________________________
>> moosefs-users mailing list
>> moo...@li...
>> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>
>

Re: [MooseFS-Users] Slow replication between internal disks of 1 chunkserver

[Davies L. <dav...@gm...>]

You can have a script to copy some files into new disk and create a
symlink for copied files.

Once the disks got balanced, you can stop the chunk server, REMOVE all
symlinks, then start chunk server.

On Thu, Apr 2, 2015 at 11:30 PM, Neddy, NH. Nam <na...@nd...> wrote:
> Hi,
>
> One disk was added to a chunkserver (2.0.60 CE) recently, but the
> performance seems to be not good compared to adding totally new
> chunkserver.
>
> All the disks are the same capacity but different brands. The
> previously new added chunkserver has performed at limit write speed of
> SATA2 (avg 60 MB/s), but the current server with brand disk added that
> only has been wrting avg. 25 MB/s. (as seen in server graphs)
>
> Does anybody have idea on this?
> Thanks,
> ~Neddy
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users



-- 
 - Davies

[MooseFS-Users] Slow replication between internal disks of 1 chunkserver

[Neddy, N. N. <na...@nd...>]

Hi,

One disk was added to a chunkserver (2.0.60 CE) recently, but the
performance seems to be not good compared to adding totally new
chunkserver.

All the disks are the same capacity but different brands. The
previously new added chunkserver has performed at limit write speed of
SATA2 (avg 60 MB/s), but the current server with brand disk added that
only has been wrting avg. 25 MB/s. (as seen in server graphs)

Does anybody have idea on this?
Thanks,
~Neddy

Re: [MooseFS-Users] input/output error in directory

[Steve W. <st...@pu...>]

On 04/02/2015 02:54 AM, Aleksander Wieliczko wrote:
> Hello everyone.
> We had similar problem with listing folders with more than one million 
> of files.
> This problem was resolved in MooseFS 2.0.60
> Which version of MooseFS are you using?
>

We're using version 2.0.50.  I mentioned earlier that I thought that 
Ubuntu clients could list the contents okay while RedHat clients 
couldn't.  After some more testing today, it looks like the only 
computer that can actually list the contents (that I've found so far) is 
the one which is also running mfs-master.

I'll need to try MooseFS 2.0.60 at some point to see if it helps get 
around the problem.

Thanks,
Steve


> Best regards
> Aleksander Wieliczko
> Technical Support Engineer
> MooseFS.com <moosefs.com>
>
> On 04/01/2015 06:06 PM, Steve Wilson wrote:
>> On 04/01/2015 11:12 AM, Ricardo J. Barberis wrote:
>>> El Miércoles 01/04/2015, Steve Wilson escribió:
>>>> Hi,
>>>>
>>>> One of our MooseFS installations has a problem with one of its
>>>> directories.  mfsdirinfo shows:
>>>>
>>>>       [root@mckinley IPAB]# mfsdirinfo database
>>>>       database:
>>>>         inodes:                    2381944
>>>>          directories:                    1
>>>>          files:                    2381943
>>>>         chunks:                    2381942
>>>>         length:                 4863750852
>>>>         size:                 168298493952
>>>>         realsize:             336596987904
>>>>
>>>>
>>>> When going into the directory and attempting to list the contents, I get
>>>> the following error from ls:
>>>>
>>>>       ls: reading directory .: Input/output error
>>>>
>>>> I can create new entries in this directory, list them, and remove them:
>>>>
>>>>       [root@mckinley x]# touch a
>>>>       [root@mckinley x]# ls -l a
>>>>       -rw-r--r-- 1 root root 0 Apr  1 10:45 a
>>>>       [root@mckinley x]# rm a
>>>>       rm: remove regular empty file `a'? y
>>>>       [root@mckinley x]#
>>>>
>>>> I don't see any MooseFS error messages in the logs and the MooseFS
>>>> filesystem check information is clean.  Any suggestions???
>>> I'm not sure, but I'd check connectivity from clients to chunkservers (i.e:
>>> firewall, routing) because creating an empty file only contacts mfsmaster and
>>> maybe ls'ing an empty file does the same.
>>>
>>> Does a simple '/bin/ls' inside database work?
>>>
>> It turns out that my RHEL 6 clients are unable to deal with the contents
>> of this directory while my Ubuntu 14.04 clients were able to access it.
>> The users who created this directory are on Ubuntu workstations but my
>> backup servers run RHEL or CentOS.  I'm fairly certain that I'm reaching
>> some OS limit since this directory has over 2 million entries in it.
>>
>> Thanks,
>> Steve
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub for all
>> things parallel software development, from weekly thought leadership blogs to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now.http://goparallel.sourceforge.net/
>>
>> _________________________________________
>> moosefs-users mailing list
>> moo...@li...
>> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>

Re: [MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Aleksander W. <ale...@mo...>]

Hello
First of all I would like to explain what these parameters mean and how
does it work:

CHUNKS_SOFT_DEL_LIMIT - Soft maximum number of chunks being
simultaneously deleted on one chunkserver (default is 10)
CHUNKS_HARD_DEL_LIMIT - Hard maximum number of chunks being
simultaneously deleted on one chunkserver (default is 25)

MooseFS mainly uses the soft delete limit. If the system is able to
remove all "to delete" chunks in one loop run, deletion limit will not
be changed.
But if the number of chunks to delete increase after a loop run,
deletion limit will be increased by factor of 1.5 but never to a value
higher than CHUNKS_HARD_DEL_LIMIT.

This mechanism was implemented to prevent MooseFS cluster from freezing
because of too many delete operations.
So if you have small number of delete operations in your system hard
limit will never be reached.

How many files are you deleting in your system?

Best regards
Aleksander Wieliczko
Technical Support Engineer
MooseFS.com <moosefs.com>

On 04/01/2015 09:48 PM, Eduardo Kellenberger wrote:
>
> To make changes to the configuration file on the parameters mfsmaster.conf
>
>  CHUNKS_SOFT_DEL_LIMIT
>  CHUNKS_HARD_DEL_LIMIT
>
>  these are not reflected in the graphs.
>
>  Anyone have idea about it.
>
>
>
>

Re: [MooseFS-Users] input/output error in directory

[Aleksander W. <ale...@mo...>]

Hello everyone.
We had similar problem with listing folders with more than one million
of files.
This problem was resolved in MooseFS 2.0.60
Which version of MooseFS are you using?

Best regards
Aleksander Wieliczko
Technical Support Engineer
MooseFS.com <moosefs.com>

On 04/01/2015 06:06 PM, Steve Wilson wrote:
> On 04/01/2015 11:12 AM, Ricardo J. Barberis wrote:
>> El Miércoles 01/04/2015, Steve Wilson escribió:
>>> Hi,
>>>
>>> One of our MooseFS installations has a problem with one of its
>>> directories.  mfsdirinfo shows:
>>>
>>>      [root@mckinley IPAB]# mfsdirinfo database
>>>      database:
>>>        inodes:                    2381944
>>>         directories:                    1
>>>         files:                    2381943
>>>        chunks:                    2381942
>>>        length:                 4863750852
>>>        size:                 168298493952
>>>        realsize:             336596987904
>>>
>>>
>>> When going into the directory and attempting to list the contents, I get
>>> the following error from ls:
>>>
>>>      ls: reading directory .: Input/output error
>>>
>>> I can create new entries in this directory, list them, and remove them:
>>>
>>>      [root@mckinley x]# touch a
>>>      [root@mckinley x]# ls -l a
>>>      -rw-r--r-- 1 root root 0 Apr  1 10:45 a
>>>      [root@mckinley x]# rm a
>>>      rm: remove regular empty file `a'? y
>>>      [root@mckinley x]#
>>>
>>> I don't see any MooseFS error messages in the logs and the MooseFS
>>> filesystem check information is clean.  Any suggestions???
>> I'm not sure, but I'd check connectivity from clients to chunkservers (i.e:
>> firewall, routing) because creating an empty file only contacts mfsmaster and
>> maybe ls'ing an empty file does the same.
>>
>> Does a simple '/bin/ls' inside database work?
>>
> It turns out that my RHEL 6 clients are unable to deal with the contents 
> of this directory while my Ubuntu 14.04 clients were able to access it.  
> The users who created this directory are on Ubuntu workstations but my 
> backup servers run RHEL or CentOS.  I'm fairly certain that I'm reaching 
> some OS limit since this directory has over 2 million entries in it.
>
> Thanks,
> Steve
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the 
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users

Re: [MooseFS-Users] Transfer stats

[Aleksander W. <ale...@mo...>]

Hi
MooseFS in "Disk" tab shows transfer speed from system - not the direct
hdd speed.
Usually OS uses write-back cache to speed up writes. As a result if
there are not so many writes then in MFS stats you see "fake" write
speed because it's only speed of copying data to kernel write-back cache.
 
Best regards
Aleksander Wieliczko
Technical Support Engineer
MooseFS.com <moosefs.com>

On 04/01/2015 12:33 PM, Neddy, NH. Nam wrote:
> My "servers" are old Intel Pentium Dual-core bpxes with SATA2 HDDs,
> but the transfer rate are enormous:
>
> https://i.imgur.com/ouN8oFs.png
>
> How did MooseFS calculated those rates?
>
> Note that all servers are connected to 1G switch, and usage is about
> 90% of 1G when there's traffic.
>
> Best regards,
> ~Nam
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the 
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users

[MooseFS-Users] CHUNKS_SOFT_DEL_LIMIT and CHUNKS_HARD_DEL_LIMIT on moosefs ce

[Eduardo K. <edu...@do...>]

To make changes to the configuration file on the parameters mfsmaster.conf

 CHUNKS_SOFT_DEL_LIMIT
 CHUNKS_HARD_DEL_LIMIT

 these are not reflected in the graphs.

 Anyone have idea about it.




-- 
-------------------------------------------------
Eduardo Kellenberger
Departamento de Infraestructura Tecnológica
DonWeb
"La actitud es todo"
Donweb.com

Nota de confidencialidad: Este mensaje y archivos adjuntos al mismo son confidenciales, de uso exclusivo para el destinatario del mismo. La divulgación y/o uso del mismo sin autorización por parte de DonWeb.com queda prohibida.
DonWeb.com no se hace responsable del mensaje por la falsificación y/o alteración del mismo.
De no ser Ud el destinatario del mismo y lo ha recibido por error, por favor, notifique al remitente y elimínelo de su sistema.

Confidentiality Note: This message and any attachments (the message) are confidential and intended solely for the addressees. Any unauthorised use or dissemination is prohibited by DonWeb.com.
DonWeb.com shall not be liable  for the message if altered or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender

Nota de Confidencialidade: Esta mensagem e seus eventuais anexos podem conter dados confidenciais ou privilegiados.
Se você os recebeu por engano ou não é um dos destinatários aos quais ela foi endereçada, por favor destrua-a e a todos os seus eventuais anexos ou copias realizadas, imediatamente.
É proibida a retenção, distribuição, divulgação ou utilização de quaisquer informações aqui contidas.

Por favor, informenos sobre o recebimento indevido desta mensagem, retornando-a para o autor.

Re: [MooseFS-Users] input/output error in directory

[Steve W. <st...@pu...>]

On 04/01/2015 11:12 AM, Ricardo J. Barberis wrote:
> El Miércoles 01/04/2015, Steve Wilson escribió:
>> Hi,
>>
>> One of our MooseFS installations has a problem with one of its
>> directories.  mfsdirinfo shows:
>>
>>      [root@mckinley IPAB]# mfsdirinfo database
>>      database:
>>        inodes:                    2381944
>>         directories:                    1
>>         files:                    2381943
>>        chunks:                    2381942
>>        length:                 4863750852
>>        size:                 168298493952
>>        realsize:             336596987904
>>
>>
>> When going into the directory and attempting to list the contents, I get
>> the following error from ls:
>>
>>      ls: reading directory .: Input/output error
>>
>> I can create new entries in this directory, list them, and remove them:
>>
>>      [root@mckinley x]# touch a
>>      [root@mckinley x]# ls -l a
>>      -rw-r--r-- 1 root root 0 Apr  1 10:45 a
>>      [root@mckinley x]# rm a
>>      rm: remove regular empty file `a'? y
>>      [root@mckinley x]#
>>
>> I don't see any MooseFS error messages in the logs and the MooseFS
>> filesystem check information is clean.  Any suggestions???
> I'm not sure, but I'd check connectivity from clients to chunkservers (i.e:
> firewall, routing) because creating an empty file only contacts mfsmaster and
> maybe ls'ing an empty file does the same.
>
> Does a simple '/bin/ls' inside database work?
>

It turns out that my RHEL 6 clients are unable to deal with the contents 
of this directory while my Ubuntu 14.04 clients were able to access it.  
The users who created this directory are on Ubuntu workstations but my 
backup servers run RHEL or CentOS.  I'm fairly certain that I'm reaching 
some OS limit since this directory has over 2 million entries in it.

Thanks,
Steve

Re: [MooseFS-Users] input/output error in directory

[Ricardo J. B. <ric...@do...>]

El Miércoles 01/04/2015, Steve Wilson escribió:
> Hi,
>
> One of our MooseFS installations has a problem with one of its
> directories.  mfsdirinfo shows:
>
>     [root@mckinley IPAB]# mfsdirinfo database
>     database:
>       inodes:                    2381944
>        directories:                    1
>        files:                    2381943
>       chunks:                    2381942
>       length:                 4863750852
>       size:                 168298493952
>       realsize:             336596987904
>
>
> When going into the directory and attempting to list the contents, I get
> the following error from ls:
>
>     ls: reading directory .: Input/output error
>
> I can create new entries in this directory, list them, and remove them:
>
>     [root@mckinley x]# touch a
>     [root@mckinley x]# ls -l a
>     -rw-r--r-- 1 root root 0 Apr  1 10:45 a
>     [root@mckinley x]# rm a
>     rm: remove regular empty file `a'? y
>     [root@mckinley x]#
>
> I don't see any MooseFS error messages in the logs and the MooseFS
> filesystem check information is clean.  Any suggestions???

I'm not sure, but I'd check connectivity from clients to chunkservers (i.e: 
firewall, routing) because creating an empty file only contacts mfsmaster and 
maybe ls'ing an empty file does the same.

Does a simple '/bin/ls' inside database work?

> Thanks!
>
> Steve

Regards,
-- 
Ricardo J. Barberis
Senior SysAdmin / IT Architect
DonWeb
La Actitud Es Todo
www.DonWeb.com
  _____

[MooseFS-Users] input/output error in directory

[Steve W. <st...@pu...>]

Hi,

One of our MooseFS installations has a problem with one of its 
directories.  mfsdirinfo shows:

    [root@mckinley IPAB]# mfsdirinfo database
    database:
      inodes:                    2381944
       directories:                    1
       files:                    2381943
      chunks:                    2381942
      length:                 4863750852
      size:                 168298493952
      realsize:             336596987904


When going into the directory and attempting to list the contents, I get 
the following error from ls:

    ls: reading directory .: Input/output error

I can create new entries in this directory, list them, and remove them:

    [root@mckinley x]# touch a
    [root@mckinley x]# ls -l a
    -rw-r--r-- 1 root root 0 Apr  1 10:45 a
    [root@mckinley x]# rm a
    rm: remove regular empty file `a'? y
    [root@mckinley x]#

I don't see any MooseFS error messages in the logs and the MooseFS 
filesystem check information is clean.  Any suggestions???

Thanks!

Steve

[MooseFS-Users] Transfer stats

[Neddy, N. N. <na...@nd...>]

My "servers" are old Intel Pentium Dual-core bpxes with SATA2 HDDs,
but the transfer rate are enormous:

https://i.imgur.com/ouN8oFs.png

How did MooseFS calculated those rates?

Note that all servers are connected to 1G switch, and usage is about
90% of 1G when there's traffic.

Best regards,
~Nam

Re: [MooseFS-Users] >>: Native windows client

[Alexey S. <al...@si...>]

How to avoid this restart?
 



-- 

С уважением, Алексей Силк
With best regards, Aleksey Silk
+7 (981) 849-12-36
skype - rootiks

-----Исходное сообщение-----
От: "Neddy, NH. Nam" <na...@nd...>
Отправлено: ‎31.‎03.‎2015 17:56
Кому: "Alexey Silk" <al...@si...>
Копия: "moo...@li..." <moo...@li...>
Тема: Re: [MooseFS-Users] >>: Native windows client

HI,

I'm running a similar scenario, that looked like: MooseFS ---- SAMBA
------ LAN.

My procedure looks like:

1. install moosefs cluster up and running (currently on Debian Wheezy)
1.i. configure mfsexports (plan for client mounting points)

2. set up a samba server, samba users, and share folders the same as
previous step

3. mfsmount and start samba service

It works but with any small changes, I have to restart samba service,
unmount/remount on windows clients. This could be awful when there are
many users, not only the possibly bottleneck if the filesystem size
increases and there are too many small files as in my case.

Best regards.
~Nedd

On Tue, Mar 31, 2015 at 12:31 PM, Alexey Silk <al...@si...> wrote:
> It will be a great example how to use samba as gateway.
> Please do it.
>
>
>
> --
>
> С уважением, Алексей Силк
> With best regards, Aleksey Silk
> +7 (981) 849-12-36
> skype - rootiks
> ________________________________
> От: Krzysztof Kielak
> Отправлено: ‎31.‎03.‎2015 8:06
> Кому: web user
> Копия: moo...@li...
> Тема: Re: [MooseFS-Users] Native windows client
>
> Dear Vineet,
>
> One option for using MooseFS on Windows is to setup a gateway through Samba
> servers possibly running on few chunkservers to spread the load from the
> clients. If you need any assistance with such setup we would be happy to
> help you.
>
> LizardFS native Windows client seems to be only “youtube-ware”.
>
> On 30 Mar 2015, at 19:55, web user <web...@gm...> wrote:
>
> Hi Guys,
>
> I got the client working on Mac and Linux. Both work great. Any chance to
> get a client working on Windows? I see that LizardFS has a native windows
> client, but about 15 minutes of googling did not get me anything.
>
> Any pointers would be much appreciated.
>
> Regards,
>
> Vineet
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now.
> http://goparallel.sourceforge.net/_______________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>
>
> Best Regards,
> Krzysztof Kielak
> Director of Operations and Customer Support
> Mobile: +48 601 476 440
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>

Re: [MooseFS-Users] >>: Native windows client

[Neddy, N. N. <na...@nd...>]

HI,

I'm running a similar scenario, that looked like: MooseFS ---- SAMBA
------ LAN.

My procedure looks like:

1. install moosefs cluster up and running (currently on Debian Wheezy)
1.i. configure mfsexports (plan for client mounting points)

2. set up a samba server, samba users, and share folders the same as
previous step

3. mfsmount and start samba service

It works but with any small changes, I have to restart samba service,
unmount/remount on windows clients. This could be awful when there are
many users, not only the possibly bottleneck if the filesystem size
increases and there are too many small files as in my case.

Best regards.
~Nedd

On Tue, Mar 31, 2015 at 12:31 PM, Alexey Silk <al...@si...> wrote:
> It will be a great example how to use samba as gateway.
> Please do it.
>
>
>
> --
>
> С уважением, Алексей Силк
> With best regards, Aleksey Silk
> +7 (981) 849-12-36
> skype - rootiks
> ________________________________
> От: Krzysztof Kielak
> Отправлено: ‎31.‎03.‎2015 8:06
> Кому: web user
> Копия: moo...@li...
> Тема: Re: [MooseFS-Users] Native windows client
>
> Dear Vineet,
>
> One option for using MooseFS on Windows is to setup a gateway through Samba
> servers possibly running on few chunkservers to spread the load from the
> clients. If you need any assistance with such setup we would be happy to
> help you.
>
> LizardFS native Windows client seems to be only “youtube-ware”.
>
> On 30 Mar 2015, at 19:55, web user <web...@gm...> wrote:
>
> Hi Guys,
>
> I got the client working on Mac and Linux. Both work great. Any chance to
> get a client working on Windows? I see that LizardFS has a native windows
> client, but about 15 minutes of googling did not get me anything.
>
> Any pointers would be much appreciated.
>
> Regards,
>
> Vineet
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now.
> http://goparallel.sourceforge.net/_______________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>
>
> Best Regards,
> Krzysztof Kielak
> Director of Operations and Customer Support
> Mobile: +48 601 476 440
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
> _________________________________________
> moosefs-users mailing list
> moo...@li...
> https://lists.sourceforge.net/lists/listinfo/moosefs-users
>

[MooseFS-Users] >>: Native windows client

[Alexey S. <al...@si...>]

It will be a great example how to use samba as gateway.
Please do it.



-- 

С уважением, Алексей Силк
With best regards, Aleksey Silk
+7 (981) 849-12-36
skype - rootiks

-----Исходное сообщение-----
От: "Krzysztof Kielak" <krz...@mo...>
Отправлено: ‎31.‎03.‎2015 8:06
Кому: "web user" <web...@gm...>
Копия: "moo...@li..." <moo...@li...>
Тема: Re: [MooseFS-Users] Native windows client

Dear Vineet,


One option for using MooseFS on Windows is to setup a gateway through Samba servers possibly running on few chunkservers to spread the load from the clients. If you need any assistance with such setup we would be happy to help you.


LizardFS native Windows client seems to be only “youtube-ware”.


On 30 Mar 2015, at 19:55, web user <web...@gm...> wrote:


Hi Guys,


I got the client working on Mac and Linux. Both work great. Any chance to get a client working on Windows? I see that LizardFS has a native windows client, but about 15 minutes of googling did not get me anything. 


Any pointers would be much appreciated. 


Regards,


Vineet


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/_______________________________________________
moosefs-users mailing list
moo...@li...
https://lists.sourceforge.net/lists/listinfo/moosefs-users



Best Regards,
Krzysztof Kielak
Director of Operations and Customer Support
Mobile: +48 601 476 440