Docstoc

Gluster_FS_Release_Notes_3.2.5

Document Sample
Gluster_FS_Release_Notes_3.2.5 Powered By Docstoc
					                         Gluster File System
                              Release Notes




Version: 3.2.5
Release: November 2011
         ds




                                                     License
         The License information is available at: www.redhat.com/licenses/rhel_rha_eula.html.




Gluster File System                            Release_Notes_3.2.5_01_A                         Pg No. 2
         ds




                                                        Table of Contents
         1.    Overview ........................................................................................................... 4
                 1.1. Product Background .................................................................................... 4
                 1.2. Product License.......................................................................................... 4
                 1.3. Providing Feedback ..................................................................................... 4
         2.    What is New in this Release? .................................................................................... 5
         3.    Dropped Features ................................................................................................. 6
         4.    System Requirements ............................................................................................ 7
                 4.1. Storage Severs ........................................................................................... 7
                 4.2. Networking Requirements ............................................................................. 7
                 4.3. Operating System Requirements ..................................................................... 7
                 4.4. File System Requirements ............................................................................. 7
         5.    Downloading and Installing GlusterFS 3.2.5 .................................................................. 8
                 5.1. Downloading GlusterFS 3.2.5 .......................................................................... 8
                 5.2. New Installation ......................................................................................... 8
                 5.3. Compatibility ............................................................................................ 8
                 5.4. Upgrade ................................................................................................... 8
                        5.4.1.    Upgrade from GlusterFS v3.2.x or 3.1.x to GlusterFS v3.2.5 ......................... 8
                        5.4.2.    Upgrade from 3.0.x to 3.2.x ................................................................ 9
         6.    Issues Resolved in GlusterFS 3.2.5 ............................................................................ 10
         7.    Known Issues ...................................................................................................... 11
         8.    Documentation ................................................................................................... 13
         9.    Appendix – Bugs Fixed ........................................................................................... 14




Gluster File System                                      Release_Notes_3.2.5_01_A                                                   Pg No. 3
         ds




         1. Overview
         This Release Notes includes the following sections for the 3.2.5 release of the GlusterFS:
              What is New in this Release
              System Requirements
              Downloading and Installing GlusterFS 3.2.5
              Issues Resolved in GlusterFS 3.2.5
              Known Issues
         Note: It is recommended that you must thoroughly review this release notes prior to installing or
         upgrading GlusterFS.

         1.1. Product Background
         GlusterFS is an open source, clustered file system capable of scaling to several petabytes and
         handling thousands of clients. GlusterFS clusters together storage building blocks over Infiniband
         RDMA or TCP/IP interconnect, aggregating disk and memory resources and managing data in a single
         global namespace. GlusterFS is based on a stackable user space design and can deliver exceptional
         performance for diverse workloads.
         The GlusterFS modular architecture allows administrators to stack modules to match user
         requirements, as needed. For example, administrators can use GlusterFS to quickly configure a
         standalone server system and later expand the system as needs grow.

         1.2. Product License
         GlusterFS 3.2.5 is licensed under the GNU General Public License v3 or later. The license for
         GlusterFS 3.2.5 can be viewed in License section.

         1.3. Providing Feedback
         Red Hat welcomes your comments and suggestions on the quality and usefulness of its
         documentation. If you find any errors or have any other suggestions, write to us at
         docfeedback@gluster.com for clarification by providing the chapter, section, and page number, if
         available.
         Red Hat offers a range of resources related to GlusterFS software:
              Discuss technical problems and solutions on the Discussion Forum
              (http://community.gluster.org)
              Get hands-on step-by-step tutorials
              (http://www.gluster.com/community/documentation/index.php/Main_Page)
              Reach Support (http://www.gluster.com/services/)




Gluster File System                                Release_Notes_3.2.5_01_A                                   Pg No. 4
         ds




         2. What is New in this Release?
         GlusterFS 3.2.5 release includes the following new feature:
              GFID-self-heal: If there are any gfid mismatches, they get automatically self-healed now.




Gluster File System                              Release_Notes_3.2.5_01_A                                 Pg No. 5
         ds




         3. Dropped Features
         No feature is dropped from this release.




Gluster File System                             Release_Notes_3.2.5_01_A   Pg No. 6
         ds




         4. System Requirements
         4.1. Storage Severs
         The following storage servers are supported:
              Processor: Intel/AMD x86 64-bit
              Disk: 8GB minimum using direct-attached-storage, RAID, Amazon EBS, and FC/Infiniband/iSCSI
              SAN disk backend using SATA/SAS/FC disks
              Memory: 1GB minimum

         4.2. Networking Requirements
         The following are the supported networks:
              Gigabit Ethernet
              10 Gigabit Ethernet
              InfiniBand

         4.3. Operating System Requirements
         GlusterFS has been tested extensively on the following environment:

           Operating System                      Architecture              Modes
           RHEL 6.1                              x86_64                    Client & Server

         GlusterFS works with other common Linux distribution like CentOS 5.1 or higher, Ubuntu 8.04 or
         higher, and Fedora 11 or higher, but has not been tested extensively.
         Packages
         Ensure that the following packages are installed:

              Bison
              Automake/ Autoconf
              Flex
              libtool
              gcc
              Portmapper (for NFS)
              Fuse
              Readline

         4.4. File System Requirements
         Red Hat recommends XFS when formatting the disk sub-system. XFS supports metadata journaling,
         which facilitates quicker crash recovery. The XFS file system can also be de-fragmented and enlarged
         while mounted and active. Any other POSIX compliant disk file system, such as Ext3, Ext4, ReiserFS
         may also work, but has not been tested widely.




Gluster File System                             Release_Notes_3.2.5_01_A                                    Pg No. 7
         ds




         5. Downloading and Installing GlusterFS 3.2.5
         You can download and install the GlusterFS 3.2.5 or upgrade to latest version.

         5.1. Downloading GlusterFS 3.2.5
         You can download the latest software to each server in your cluster from
         http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/.

         5.2. New Installation
         The installation process for GlusterFS server is available at:
         http://www.gluster.com/community/documentation/index.php/Gluster_3.2:_Installing_GlusterFS_S
         erver.

         5.3. Compatibility
         Release 3.2.5 of GlusterFS is compatible with 3.2.1 or higher releases of GlusterFS.
         Release 3.2.5 of GlusterFS is not compatible with 2.0.x, 3.0.x, 3.1.x, and 3.2.0 releases of GlusterFS.

         5.4. Upgrade
         Red Hat recommends that you back up your data before upgrading to GlusterFS 3.2.5.
         NOTE: Configurations generated outside the scope of gluster CLI are neither recommended nor
         supported by Red Hat.

         5.4.1. Upgrade from GlusterFS v3.2.x or 3.1.x to GlusterFS v3.2.5
         In an environment with replicated bricks it is recommended that you upgrade a single storage server,
         confirm stability, and then upgrade the replica.

         Use the same installation method for the upgrade as the original glusterfs installation.

         Using RPMs on RHEL, CentOS, Fedora

         1. Download the 3.2.5 RPMs from

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/RHEL/
              or

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/CentOS/
              or

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/Fedora/


         2. Run rpm using the following command:

              # rpm -U glusterfs*




Gluster File System                               Release_Notes_3.2.5_01_A                                      Pg No. 8
         ds




         Using dpkg on Debian, Ubuntu

         1. Download the 3.2.5 packages from

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/Debian/

              or

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/Ubuntu/
         2. Run dpkg using the following command:

              # dpkg -i glusterfs*

         Building from source

         1. Download the 3.2.5 source code from

              http://download.gluster.com/pub/gluster/glusterfs/3.2/3.2.5/glusterfs-3.2.5.tar.gz

         2. Unpack and install GlusterFS using the following commands:

              # gunzip glusterfs-3.2.5.tar.gz

              # tar xvf glusterfs-3.2.5.tar

              # cd glusterfs-3.2.5

              # ./configure

              # make

              # make install

         3. Stop and start GlusterFS using the following commands, this step will disconnect Gluster Native
            clients.

              # killall glusterfsd

              # killall glusterfs

              # killall glusterd

         4. Start GlusterFS using the following command:

              # /etc/init.d/glusterd start

         5.4.2. Upgrade from 3.0.x to 3.2.x
         For detailed instructions on upgrading from 3.0.x to 3.2.x, see
         http://www.gluster.com/community/documentation/index.php/Gluster_3.0_to_3.2_Upgrade_Guide.




Gluster File System                             Release_Notes_3.2.5_01_A                                      Pg No. 9
         ds




         6. Issues Resolved in GlusterFS 3.2.5
         The following major issues are resolved in this release:
              Bug 3734 - Fixed the issue of gfid mismatches and now they get automatically self-healed.
              Bug 3576 – Fixed the issue of excessive memory utilization after rebalance operation.
              Bug 3636 – Fixed the issue of synctask_wrap() crashing.
              Bug 3711 – Fixed the issue of glusterd log showing different file sizes after rebalance operation.
              Bug 3747 – Fixed the issue of glusterfs client crashing because of gfid being NULL.
              Bug 3774 – Fixed the issue of glusterfs build failure over stripe.




Gluster File System                                Release_Notes_3.2.5_01_A                                        Pg No. 10
         ds




         7. Known Issues
         The following are the known issues:
              Issues related to Distributed Replicated Volumes:
                     When process has done ‘cd’ into a directory, stat of deleted file recreates it (directory self-
                      heal not triggered).
              In GlusterFS replicated setup, if you are inside a directory (for example, ‘Test’ directory) of
              replicated volume. From another node, you will delete a file inside ‘Test’ directory. Then if you
              perform 'stat' operation on the same file name, the file will be automatically created (that is, a
              proper directory self-heal is not triggered when process has done ‘cd' into a path).
                     Open fd self-heal blocks the I/O on fd.
              While doing self-heal on open file descriptors in replicate, the I/O operations on that particular
              file descriptor may get blocked.
              Issues related to Distributed Volumes:
                     Rebalance does not happen if bricks are down.
              Currently while running rebalance, make sure all the bricks are in operating or connected state.
                     Rebalance can happen to already filled subvolume.
              Current algorithm of rebalance is not considering the free-space in the target brick before
              migrating data. This enhancement is under development and will be available shortly.
              There may be minor I/O glitches when Rebalance operation is performed. The live rebalance
              feature will be available in upcoming in 3.3.x releases. It is recommended to perform rebalance
              operation when there are no critical IO operations are happening.
              glusterfsd - Error return code is not proper after daemonizing the process.
              Due to this, scripts that mount glusterfs or start glusterfs process must not depend on its return
              value.
              gluster NFS server - Any operations on the files that exists beyond a directory depth of 16, users
              may experience intermittent ESTALE (Stale NFS handle) errors. For example, my-filename
              located in the path:
              /nfs/mount/point/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/my-filename
              glusterd - Parallel rebalance
              With the current rebalance mechanism, the machine issuing the rebalance is becoming a
              bottleneck as all the data migrations are happening through that machine.
              Parallel operations (add brick, remove brick, and so on) with CLI from different nodes can crash
              glusterd.
              After # gluster volume replace-brick VOLNAME Brick New-Brick commit command
              is issued, the file system operations on that particular volume, which are in transit will fail.
              Command # gluster volume replace-brick … will fail in a RDMA set up.
              If files and directories have different GFIDs on different backends, GlusterFS client may hang or
              display errors.

              Work Around: The workaround for this issue is explained at:
              http://gluster.org/pipermail/gluster-users/2011-July/008215.html
              Downgrading from 3.2.x to 3.1.x




Gluster File System                                   Release_Notes_3.2.5_01_A                                          Pg No. 11
         ds


              If you are using 3.2.x, then the new features are enabled in the default volume files (i.e. new
              translators). So after the downgrade, old versions fail to understand the new options/ translators
              and fail to start.
              Work Around: Before starting downgrade procedure, run the following commands:
              # gluster volume reset <VOLNAME> force
              # gluster volume geo-replication stop <MASTER> <SLAVE>
              Now you can downgrade to 3.1.x.
              Run any parameter changing operations on the volume. For example, operations like # gluster
              volume set <VOL> read-ahead off and # gluster volume set <VOL> read-ahead
              on.
              Issues related to Directory Quota:
                     Some writes can appear to pass even though the quota limit is exceeded (write returns
                      success). This is because they could be cached in write-behind. However disk-space would
                      not exceed the quota limit, since when writes to backend happen, quota does not allow
                      them. Hence it is advised that applications should check for return value of close call.
                     If a user has done ‘cd' into a directory on which the administrator is setting the limit, even
                      though the command succeeds and the new limit value will be applicable to all the users
                      except for those users’ who has done ‘cd' in to that particular directory. The old limit value
                      will be applicable until the user has ‘cd’ out of that directory.
                     Rename operation (that is, removing oldpath and creating newpath) requires additional disk
                      space equal to file size. This is because, during rename, it subtracts the size on oldpath after
                      rename operation is performed, but it checks whether quota limit is exceeded on parents of
                      newfile before rename operation.
                     With striped volumes, Quota feature is not available.
              Issues related to POSIX ACLs:
                     Even though POSIX ACLs are set on the file or directory, the ‘+’ (plus) sign in the file
                      permissions will not be displayed. This is for performance optimization and will be fixed in a
                      future release.
                     When glusterfs is mounted with "-o acl", directory read performance can be bad. Commands
                      like recursive directory listing can be slower than normal.
                     When POSIX ACLs are set and multiple NFS clients are used, there could be inconsistency in
                      the way ACLs are applied due to attribute caching in NFS. For a consistent view of POSIX ACLs
                      in a multiple client setup, use ‘-o noac’ option on NFS mount to switch off attribute caching.
                      This could have a performance impact on operations involving attributes.
              The following are few known missing (minor) features:
                     locks - 'mandatory' locking is not supported.
                     NLM (Network Lock Manager) is not supported.




Gluster File System                                   Release_Notes_3.2.5_01_A                                         Pg No. 12
         ds




         8. Documentation
         Product documentation of GlusterFS is available at:
         http://www.gluster.com/community/documentation/index.php/Main_Page .




Gluster File System                             Release_Notes_3.2.5_01_A        Pg No. 13
         ds




         9. Appendix – Bugs Fixed
         The following are the list of bugs fixed in this release:
             Bug             Bug Title
                                                                             URL
           Number
                      Log message enhancements in              http://bugs.gluster.com/show_bug.cgi?id=2346
          2346
                      GlusterFS - phase 1
          2550        Build warnings                           http://bugs.gluster.com/show_bug.cgi?id=2550
                      Read child is set without checking       http://bugs.gluster.com/show_bug.cgi?id=2586
          2586
                      the xattr
                      Directories changing to d---------       http://bugs.gluster.com/show_bug.cgi?id=2591
          2591
                      permission after trying to delete
                      Make geo-replication package             http://bugs.gluster.com/show_bug.cgi?id=2744
          2744
                      friendly
                      If gsyncd is spawned by glusterd, it     http://bugs.gluster.com/show_bug.cgi?id=2787
          2787
                      ignores signal
                      Volume create succeeds even though       http://bugs.gluster.com/show_bug.cgi?id=2803
                      there is a volume with an export
          2803
                      directory in the path of current
                      volume
                      Make volume top command exit             http://bugs.gluster.com/show_bug.cgi?id=2882
          2882
                      status as zero on success
                      Make 'Profile info' output more self-    http://bugs.gluster.com/show_bug.cgi?id=3028
          3028
                      explanatory
                      inode number should be constant on       http://bugs.gluster.com/show_bug.cgi?id=3042
          3042
                      storage
          3234        gerrit integration fixes                 http://bugs.gluster.com/show_bug.cgi?id=3234
                      afr-open.c reporting warning message     http://bugs.gluster.com/show_bug.cgi?id=3443
          3443        "None of the files are open" even
                      though the files are already opened
          3446        Solaris build failures in cli1-xdr.c     http://bugs.gluster.com/show_bug.cgi?id=3446
                      rpc: Need to take a ref on rpc_clnt      http://bugs.gluster.com/show_bug.cgi?id=3511
          3511        before unwinding saved frames and
                      call_bail.
                      Option to rotate the geo-replication     http://bugs.gluster.com/show_bug.cgi?id=3519
          3519
                      logs.
                      Self-heal of special files (fifo,        http://bugs.gluster.com/show_bug.cgi?id=3525
          3525        character, block, and symlinks) does
                      not correctly reset trusted.afr xattrs
                      In nfs, socket write failures not        http://bugs.gluster.com/show_bug.cgi?id=3526
          3526
                      handled properly
                      'profile info' does not fetch the        http://bugs.gluster.com/show_bug.cgi?id=3553
          3553        statistics from other nodes in the
                      cluster
                      Nfs crashes if file name is 255          http://bugs.gluster.com/show_bug.cgi?id=3572
          3572
                      characters or more
                      Rebalance on Distributed-Replicate in    http://bugs.gluster.com/show_bug.cgi?id=3576
          3576        fails and consumes more space then
                      before
                      mem-pool creation for rb-tree hash       http://bugs.gluster.com/show_bug.cgi?id=3608
          3608
                      entries into init.




Gluster File System                               Release_Notes_3.2.5_01_A                                    Pg No. 14
         ds


             Bug             Bug Title
                                                                              URL
           Number
                      marker shows a lot of "invalid            http://bugs.gluster.com/show_bug.cgi?id=3609
          3609
                      argument:inode" errors
          3636        glustershd crashd in synctask_wrap        http://bugs.gluster.com/show_bug.cgi?id=3636
          3663        pic code hangs after fix for bug 3505     http://bugs.gluster.com/show_bug.cgi?id=3663
                      Cannot add iobuf into iobref during       http://bugs.gluster.com/show_bug.cgi?id=3679
          3679
                      mmap test
                      [glusterfs-3.3.0qa14] glusterd            http://bugs.gluster.com/show_bug.cgi?id=3700
          3700        crashed when volume stop command
                      is issued
          3709        Volume replace-brick unstable             http://bugs.gluster.com/show_bug.cgi?id=3709
                      glusterd log shows "file sizes are        http://bugs.gluster.com/show_bug.cgi?id=3711
          3711        different" after rebalance is
                      performed
                      Present 'volume top write-perf' and       http://bugs.gluster.com/show_bug.cgi?id=3719
          3719        'read-perf' output in more cleaner
                      way
                      Extended attributes are not               http://bugs.gluster.com/show_bug.cgi?id=3721
          3721
                      preserved after rebalance
          3727        GFID mismatch on file                     http://bugs.gluster.com/show_bug.cgi?id=3727
                      Rebalance CLI command should not          http://bugs.gluster.com/show_bug.cgi?id=3729
          3729
                      work if there is no distribute involved
                      Backport the gfid self-heal patch to      http://bugs.gluster.com/show_bug.cgi?id=3734
          3734
                      3.2
          3737        Remove EXCEPTIONS file                    http://bugs.gluster.com/show_bug.cgi?id=3737
                      Add-brick gives confusing error           http://bugs.gluster.com/show_bug.cgi?id=3738
          3738        message when it fails during replace-
                      brick
                      glusterfs client crashed because of       http://bugs.gluster.com/show_bug.cgi?id=3747
          3747
                      gfid being NULL
                      Deadlock between                          http://bugs.gluster.com/show_bug.cgi?id=3753
          3753        pl_update_refkeeper and process
                      state dump
                      gnfs process crashed while running        http://bugs.gluster.com/show_bug.cgi?id=3757
          3757
                      dbench on striped volume
          3760        gfid differs on different subvolumes      http://bugs.gluster.com/show_bug.cgi?id=3760
          3761        Change in log level to `lock' messages    http://bugs.gluster.com/show_bug.cgi?id=3761
                      dbench fails with 'Reply submission       http://bugs.gluster.com/show_bug.cgi?id=3765
          3765
                      failed' error
                      TestCase : "Lookup on files/dirs          http://bugs.gluster.com/show_bug.cgi?id=3767
                      which doesn't have gfid's when any of
          3767
                      the brick in replicate volume is down
                      is not allowed " fails
                      Replicate asserted due to gfid being      http://bugs.gluster.com/show_bug.cgi?id=3770
          3770
                      NULL
          3774        glusterfs build fails over stripe         http://bugs.gluster.com/show_bug.cgi?id=3774
                      glusterfs client crashed since gfid       http://bugs.gluster.com/show_bug.cgi?id=3783
          3783
                      was NULL
          3730        valgrind says Invalid reads               http://bugs.gluster.com/show_bug.cgi?id=3730
                      Parallel untar/rm followed by a RM        http://bugs.gluster.com/show_bug.cgi?id=3793
          3793
                      fails with ENOTEMPTY




Gluster File System                                Release_Notes_3.2.5_01_A                                    Pg No. 15
         ds


             Bug             Bug Title
                                                                              URL
           Number
                      Split-brain/all-fool xattr directory      http://bugs.gluster.com/show_bug.cgi?id=3796
          3796        self-heal should try to merge the
                      entries




Gluster File System                                Release_Notes_3.2.5_01_A                                    Pg No. 16

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:6
posted:2/17/2012
language:
pages:16