FAQs from MailingList
From GlusterDocumentation
Currently this page is under construction. Mean while you can just have a look on mailing list archives before posting your questions directly to Mailing-List.
Contents |
1
Q > So the question is, what makes this different from running two instances on different ports? What kind of check is being done in glusterfsd and what are the reasons that this should be prevented?
A You can run multiple glusterfsd instances on the same machine. the PID file is only a mechanism which helps you write init.d scripts neatly. things you need to take care -
1. if you are specifying pidfile, do not specify the same path for multiple daemons
2. do not export the same backend directory with different glusterfsd daemons.
3. take care to see to that both do not try to bind to the same tcp port etc.
apart from that, the two daemons are oblivious to each other and they can have same volume names defined as well.
Make sure client and server runs same version of GlusterFS.
2
Q: I get following errors when trying to compile GlusterFS? whats wrong?
../../libglusterfs/src/.libs/libglusterfs.so: undefined reference to `yyin' ../../libglusterfs/src/.libs/libglusterfs.so: undefined reference to `yylex' collect2: ld returned 1 exit status make[2]: *** [glusterfsd] Error 1
A: Just make sure you have flex and bison packages installed in the system.
3
Q I think I did something wrong, but couldn't able to figure out whats it. Can you help me?
bash# df -h | grep glusterfs3 glusterfs 56G 186M 53G 1% /mnt/glusterfs3 bash# touch /mnt/glusterfs3/foo touch: cannot touch `/mnt/glusterfs3/foo': Transport endpoint is not connected bash# ls -l /mnt/glusterfs3/ ls: /mnt/glusterfs3/: No such file or directory
'df' output is proper as in it shows proper sum of the exported partitions.
A Looking at the behavior, it looks like, unify's namespace is not properly functioning, or may be offline as of now.
General FAQs
- How would you back up data to tape?
GlusterFS mounts appears as a regular POSIX compliant disk file system. Backup procedure is same as backing up your local disk file system or your NFS mounts.
- Can you use traditional back up software like Veritas NetBackup?
Yes.
- How does the system recover from a failed disk?
GlusterFS has self-heal functionality to fix inconsistencies automatically while the file system is mounted and actively serving data. If Automatic File Replication (AFR) feature is used, then the applications won't even notice the disk or total brick failure. When a failed brick or disk comes back online, GlusterFS will automatically self heal itself.
Without AFR, GlusterFS relies on the underlying RAID controller and disk filesytem for data protection.
- What are the RAID levels supported and recommended?
At the disk level, RAID6/RAID5/RAID10 is good. GlusterFS itself doesn't enforce or prefer any type of RAID protection. It depends on the data and the applications. In fact, for video streaming, we recommend simple JBOD mode or direct multi-port SAS controller.
GlusterFS also has file level RAID-0, RAID-1, RAID-10 like functionalities spanning across the storage bricks.
- How do you move data from one storage brick to another ?
Typically we use tar or cpio in conjunction with scp or ftp to move the files from one brick to another. It is also possible to simply comment out the brick volume and re-distribute the files automatically by simply copying the files back to the mounted volume.
In 1.4, you will be able to hot-add/suspend/remove through the web interface.
- Is there a way to tier data?
No. Hierarchical Storage Manager (HSM) functionality is planned for 1.5 release.
- How do you access data in a Windows environment? CIFS?
GlusterFS volumes are exported to MS Windows environment through Samba server. Samba supports SMB and CIFS clients.
- How do file and share level permissions work?
Works like a standard windows file server. Samba layer translates MS Windows file systems permissions to Unix POSIX uid/gid/rwx-rwx-rwx attributes transparently.
- Does it integrate with Windows 2003 Active Directory?
GlusterFS uses Samba server for Microsoft Windows file sharing. Samba has active directory support.
- What are the maximum files per volume?
1.3.x supports up to 2^57 (144115188075855872) files in a single volume. 1.4 will support 2^64 (18446744073709551616) files.
- What are the maximum inodes per volume?
Same count as above.
- How do you remove an entire ‘brick’ from the system?
We simply redistribute the files to other bricks and delete the volume declaration from the configuration file.
(From 1.4 onwards, you can dynamically add/remove/suspend storage bricks through the web interface).
- What does the metadata on each file look like? Creation date,
last access date, etc.
GlusterFS stores files and folders as it is on the back-end disk (like NFS). Even striped or mirrored files will look like regular files on disk.
- Is the glusterfs pre-installed on the bricks when purchased or do
we need to configure them when the hardware arrives?
We will install/configure GlusterFS when the hardware arrives at your facility. We are currently in discussions with tier-1 hardware suppliers to integrate GlusterFS some of their selected storage hardware models.
- How many network interfaces does each brick use for
communications (client, inter-communication)?
GlusterFS can make use of all the available ports on the brick. It is configurable.
- Does the network traffic need to be separated into their own
network vlan (again client vs interconnects)?
No separate VLAN required. Each storage brick requires an IP address and should be reachable from any client directly. When accessed using native GlusterFS, you can completely avoid bricks communicating with each other.
- Does the glusterfs auto rebalance the data load across all nodes
in the network, or is this a manual process?
Yes, it is automatic.
- Is there a way to monitor disk usage on a per brick, and disk
basis? (see if there is a hot disk or brick taking all the load)
GlusterFS provides "/proc" like virtual interface, where each brick/disks will appear like sub directories. Admin can use regular file system utilities like du, df .. and find out.
(1.4 onwards you can monitor the disk usage using web interface).
- Any performance monitoring to show throughput, and does it keep a
history?
This feature is currently in development and scheduled for 1.4 release (3 months). It keeps a history in RRD database and you can view them graphically through the web interface.
How to use GlusterFS on Windows machine
Question:
Use case - Mounting remote GlusterFS system on PC as X-drive and scheduling backups
Steps involved :
1.Install GlusterFS on Amazon EC2 instance
2.Mount remote GlusterFS on my local PC as X-drive (What command need to run for mounting as X-drive )
3.Can we do any compression and encryption of data when transferring to remote GlusterFS system ?
4. Scheduling backup using microsoft native backup scheduler.
Answer:
Typically our Windows customers are using GlusterFS over SAMBA/CIFS within their LAN/WAN environment. In order to access GlusterFS volumes over the Internet from a windows system, WebDAV or FTP type protocols are better suited.
At the top level, steps are:
- Setup GlusterFS clustered storage
Documentation: http://www.gluster.org/docs/index.php/GlusterFS
- Export GlusterFS volumes over Apache/WebDAV URL
Documentation: http://httpd.apache.org/docs/2.0/mod/mod_dav.html
- Use Microsoft Windows "Map Network Drive" facility to map the URL as a drive.
Documentation: http://msdn2.microsoft.com/en-us/library/ms881375.aspx http://learn.iis.net/page.aspx/386/using-the-webdav-redirector/
Once it is mapped as a drive, Applications can read and write files on it normally, including backup tools.
We haven't tested such a setup so far. More simplistic approach is to use a backup software that supports FTP directly.
GlusterFS can do compression and encryption, but in your case, it is better to leave this job to the backup software itself to ensure that data is encrypted and compressed before transferred accross the Internet. Most backup softwares support encryption and compression.
