Moose File System v 1.6.20 released

This release has two improvements:

  • (cs) fixed "packet too big" issue during register to master (split big register packet with all chunks info into small packets)
  • (cs, metalogger,master) added minimum socket timeout (ten seconds)

As chunkservers registration process was changed, the chunkservers in version 1.6.20 would work only with the master server in 1.6.20+ version.

 

Please first update the master server to 1.6.20 and only after that update the chunkservers.

MooseFS

2011-01-17 10:00 by Michał Borychowski MooseFS MooseFS (comments: 18)

Go back

Comment by Steve | 2012-02-12

Updates are in the pipeline, coming soon. Project is alive.

Comment by jooj | 2012-02-11

any news ? moosefs is still alive ?

Comment by liushaobo | 2012-02-10

hi， When will an updated version

Comment by Frank Han | 2011-11-22

Can you send me some background information about MooseFS ?for example,when is it first published and who or which company developed it. Thank you.

Comment by Donny Bay | 2011-09-26

Hi,

I've been experimenting with MooseFS as a way of aggregating scratch space in a compute cluster. So far it works very well, much more stable than GlusterFS and more resilient than OrangeFS.

I was wondering if MooseFS is safe for parallel I/O, in particular MPI I/O?

Also, would using an IP load balancer in front of several metadata servers each accessing a SAN filesystem with Red Hat GFS2 for the metadata logs work as a way of doing load balancing/failover?

Comment by leo | 2011-09-16

Drbd+Heartbeat+Pacemaker realization of MFS (moosefs) of high-availability

http://hi.baidu.com/leolance/blog/item/7ac035205870f020c9955905.html

Comment by Dennis McEntire | 2011-09-14

Thank you for the MooseFS system, we have been using it for about a year and it is very reliable and stable. We tried XtreemFS and MogileFS and had reliability issues for both. The Moose system is running very well and fits all our needs.

I was wondering if there were plans to make an auto failover master option? We had a master fail but the logger was able to take over and we got back up and running pretty quickly, but it was not automatic.

Thank you for your help and for the great system!

Comment by mobedu | 2011-08-02

mfsexports.cfg Does not support domain.Want to provide improved methods

Comment by Sam | 2011-07-22

Hi,

Thank you very much for this release, it is the best network file system for me ;)

When the next release?

Best regards.

Comment by sickness | 2011-06-03

To succesfully compile under openindiana (and so I suppose under solaris in general) I had to add these 3 lines on top of config.h:
#ifndef _POSIX_PTHREAD_SEMANTICS
#define _POSIX_PTHREAD_SEMANTICS
#endif
and then I had to install the solaris packages for fuse, the solaris package for gettext, and the solaris package for header-math

Comment by hanyw | 2011-04-06

hi,i want to ask you one question，please answer me,thanks!
problem one: as we know,for the moment, the clients need to use FUSE,if i use NFS, wether to support "7 mul 24 hours" model ??

Comment by guowl | 2011-03-25

hi,i want to ask for you this question，please answer me


1.Does the MooseFs Support HPUX IA64 ???

2. does the client must use fuse???

3. client must use HPUX IA64 ，what should i do????

Comment by Jofly | 2011-03-22

Can you send me some background information about MooseFS ?for example,when is it first published and who or which company developed it. Thank you.

Comment by apual | 2011-03-17

MooseFS is greate.

How can I join the work group to develop it ?

Comment by Jack | 2011-02-04

Comment by wangweichun | 2011-01-31

Comment by ZHENG | 2011-01-18

speeding :), so cool~

Comment by z1y | 2011-01-17

Thank you very much for your work. MooseFS just rocks :)

Add a comment