Issue 330 in s3fs: Multiple Mount Support

[s3...@googlecode.com]

Status: New
Owner: ----
Labels: Type-Support Priority-Medium

New issue 330 by br...@redix.com.br: Multiple Mount Support
http://code.google.com/p/s3fs/issues/detail?id=330

Detailed description of support request:
Hi,
Im making some tests with bacula backup and s3fs
The backup works fine with one test server using s3fs with use_cache option.
I have to mount multiple servers directly do s3 bucket. Its possible to
mount s3fs on several computers at the same time? like CIFS filesystem and
others
I need read/write mount for the bucket in all computers. My filesystem can
corrupt if i mount and copy files to s3 bucket at same time?




===================================================================
The following information is very important in order to help us to help
you. Omission of the following details may delay your support request or
receive no attention at all.
===================================================================
Version of s3fs being used (s3fs --version):
Amazon Simple Storage Service File System 1.64

Version of fuse being used (pkg-config --modversion fuse):
2.8.4

System information (uname -a):
2.6.18-348.3.1.el5 #1 SMP Tue Mar 5 13:19:32 EST 2013 x86_64 x86_64 x86_64
GNU/Linux

Distro (cat /etc/issue):
Red Hat Enterprise Linux Server release 5.6 (Tikanga)

s3fs command line used (if applicable):

/etc/fstab entry (if applicable):
s3fs#bucket /mnt/backups3 fuse
use_cache=/backup/tmps3,allow_other,_netdev,nosuid,nodev,url=https://s3.amazonaws.com
0 0

s3fs syslog messages (grep s3fs /var/log/syslog):
none

--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings

[s3...@googlecode.com]

Updates:
Status: Done

Comment #1 on issue 330 by ggta...@gmail.com: Multiple Mount Support
http://code.google.com/p/s3fs/issues/detail?id=330

Hi,

You can run s3fs on several servers which mounts same bucket.
If you need to run several s3fs process on one server, you have to set
cache directory for each process.

** Attention
When the file(object) is updated from some servers at same time, maybe the
file(object) is updated by lastest request.
At first, you need to know the logic that s3fs has the cache for object
stat information.
This cache logic is one of trouble that s3fs can not know the object update.
If you need to escape this trouble, you can set stat_cache_expire
option(see. man page).
As an exception, when s3fs lists objects in a directory(ex. ls command),
s3fs requests "listbucket" and this response has updated object ETag.
On this case, s3fs can know the object update because the ETag is updated
after the object is updated.

Please try to run several servers or processes, and if you find something
wrong please let me know.
Last, please use latest version(now v1.66).

Regards,