moosefs

Dear MooseFS Users,

MooseFS 3.0.96 is stable now! Please find the list of fixes this release includes below:

• (tools) fixed segfault in mfsscadmin

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.95 is stable now! Please find the list of fixes this release includes below:

• (mount+master) added update master version during reconnection to master (after updating master server)
• (master) fixed winattr field size in setattr command
• (mount) fixed dirattrcache (freezes or inefficiences during listing directories - introduced in 3.0.93)
• (cgi) fixed displaying master charts without leader (pro only)
• (mount) added info about using open dir cache in oplog/ophistory

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.94 is stable now! Please find the list of fixes this release includes below:

• (mount) fixed support for CREATE/REPLACE flags in setxattr
• (mount) added headers for flock defines
• (all) added check for poll.h header file and use it instead of sys/poll.h if possible
• (master) added test for WRITE access on directory during moving between different parents
• (master) added clearing suig/sgid during write
• (master+mount) increased max symlink path size to 4096 (posix compliance)
• (master) redesigned xattr storage (much faster and uses less memory)
• (master) improved hash map in xattr and acl (static -> dynamic)
• (cgi+cli) added possibility to define master group as set of IP adresses divided by comma or semicolon
• (cs) fixed using fsync on closed descriptors (after change FSYNC_BEFORE_CLOSE option)
• (master) fixed removing default acl
• (master+mount) added support for basic Windows attributes (hidden,ro,system,archive)
• (cgi+cli) fixed working with followers only
• (all) added api for reading config parameters from master and chunkserver
• (tools) increased timeout from 10 seconds to 20 seconds (needed for huge snapshots)

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.92 is stable now! Please find the list of fixes this release includes below:

• (master+tools) added chunk slices to mfsappendchunks
• (tools) added archive mode tools
• (master+mount) fixed getfacl (unnecessary check for read rights for uid/gid)
• (master) fixed changing acl mask during setattr

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.91 is stable now! Please find the list of fixes this release includes below:

• (all) silence false warnings generated by static analyzers (clang and cppcheck)
• (master) fixed quota testing routine used during move/rename
• (all) added README.md to distribution
• (cs+mount) decreased delay in conncache
• (mount) fixed reading done by many (20+) threads using one descriptor (premature removing requests from queue during cleaning) - resolves some issues with running VMs on MooseFS

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.90 is stable now! Please find the list of fixes this release includes below:

• (master) fixed move/rename with quota
• (master) fixed chunk counters shown in cli/cgi
• (master+tools) added option preserve hardlinks to mfsmakesnapshot
• (master) fixed ACL copying during mfsmakesnapshot
• (mount) added fixing file length in all inodes after write
• (cs) split bitfiled into separate bytes (fixed potential race condition)
• (master) setting operation to NONE before sending status (silence unnecessary messages in some cases)
• (cs) increased verbosity of crc-error messages
• (cs) fixed invalidating preserved block in case of truncate down followed by truncate up
• (mount) increased master-proxy buffer sizes

Best regards,
Peter / MooseFS Team

Dear MooseFS Users,

MooseFS 3.0.88 is stable now! Please find the list of fixes this release includes below:

• (mount) added read cache clean on write (same file access using different descriptors)
• (mount) added missing cond_destroy in readdata.c (fix sent by Jakub Ratajczak)
• (master) fixed initializing packet size for reading 'sustained' directory
• (all) fixed zassert for printing correct statuses in case of pthread functions
• (mount) fix fleng in finfo after truncate (patched by Davies Liu @davies)
• (mount) fix overlapped read (patched by Davies Liu @davies)
• (mount) fixed invalidating chunk cache after truncate
• (mount) fixed fleng handling in read worker
• (mount) fixed handling BREAK state in read worker
• (mount) changed handling data invalidation in read module (sometimes could be less efficient, but it is much more safer)
• (tools) fixed number parsing (patched by Paweł Gawroński)
• (cli) fixed printed host/port options
• (mount) moved pipes from requests to workers (read and write - huge decrease of descriptors used by mount)
• (mount) changed signal to broadcast in rwlock (fixed very rare read/write deadlock)
• (mount) fixed 'open descriptors' leak (lookup(with data for open) -> open(with cached data) -> close)
• (mount) fixed potential 'race condition' - free csdata during access
• (master) split (only internally) sustained folder into 256 subfolders (too many sustained files caused socket timeouts in master)

Best regards,
Peter / MooseFS Team

• (master) fixed leader-follower resynchronization after reconnection (pro only)
• (master) fixed rehashing condition in edge/node/chunk hashmaps (change inspired by yinjiawind)
• (mount) fixed memory leak (also efficiency) on Linux and potential segfaults on FreeBSD (negative condition)
• (mount) fixed race condition for inode value in write module
• (mount) better descriptors handling (lists of free elements, inode hash)
• (mount) better releasing descriptors on FreeBSD
• (cs) fixed time condition (patch send by yinjiawind)
• (mount) fixed logging unnecessary messages on FreeBSD (closes #19)
• (master) added condition that prevents deleting unused chunks from structures too early

• (master) fixed setting acl-default without named users or named groups
• (master) fixed master-follower synchronization after setting storage class
• (cs) changed header size from 5k to 8k (due to 4k-sector hard disks)
• (all) silenced message about using deprecated parameter 'oom_adj'
• (mount) fixed FreeBSD delayed release mechanism
• (mount) added rwlock for chunks

• (mount) fixed using connection cache in readdata module
• (freebsd) changed name in ports file from moosefs to moosefs2

• (cli) fixed displaying quota info in plain mode
• (mount) fixed code to silence some compiler warnings

• (mount) fixed oom killer disabling (setting oom_adj and oom_score_adj)
• (cli) fixed displaying inactive mounts
• (mount) added fsync before removing any locks
• (daemons) added disabling oom killer (Linux only)
• (all) small fixes in manpages
• (mount) fixed handling nonblocking lock commands (unlock and try) in both locking mechanisms
• (daemons+mount) changed default settings for limiting malloc arenas (Linux only)

• (master) fixed chunk loop (in some cases chunks from the last hash position might be left unchecked)
• (master) fixed storage class management (fixed has_***_labels fields)

• (master) fixed 'flock' (change type of lock SH->EX and EX->SH caused access to freed memory and usually SEGV)

• (cs) fixed serious error that may cause data corruption during internal rebalance (intr. in 3.0.75)

This is a first Stable release in MooseFS 3.0 branch.

With MooseFS 3.0 release we introduce great new features:

• Storage Classes for better data distribution and redundancy management,
• Global locks support compatible with POSIX ioctl() and lockf() semantics and BSD compatible flock() semantics,
• New I/O synchronisation mechanism between different mounts, for better handling concurrent I/O operations from different clients,
• ATMM (Automatic Temporary Maintenance Mode),
• Automatic File "sparsification".

... and many more.

Changes in MooseFS 3.0.77 in comparison to 3.0.76:

• (mount) added assertions after packet allocation in master communication module
• (manpages) fixes related to storage classes
• (all) improved error messages used for storage classes

• (master) fixed resolving multi path for root inode
• (master) fixed licence expiration behaviour (pro only)

• (all) fixed cppcheck warnings/errors (mostly false negatives)
• (cs) added file sparsification during chunk write (also chunk replication, chunk duplication etc.)
• (mount) fixed write data inefficiency when I/O was performed by root
• (master) fixed removing too early locked chunks from data structures
• (tools) fixed reusing local ports (connection to master returned EADDRNOTAVAIL)
• (all) changed default action from restart to start
• (all) added exiting when user defined config (option -c) cannot be loaded
• (cs) changed handling chunk filename (dynamic filename generation - cs should use less memory)
• (netdump) changed 'moose ports only' option to 'port range'
• (mount) limited number of files kept as open after close
• (cs) changed subfolder choosing algorithm
• (mount) changed mutex lock to rw-lock for I/O on the same descriptor
• (mount) added link to mfsmount from '/sbin/mount.moosefs' (Linux only)
• (all) introduced "storage classes" (new goal/labels management)
• (master+cs) introduced 'temporary maintenance mode' (automatic maintenance mode after graceful stop of cs)
• (master+mount) added fix for bug in FreeBSD kernel (kernel sends truncate before first close - FreeBSD only)
• (cs) fixed setting malloc pools on Linux

• (master+mount) added fix for bug in FreeBSD kernel (kernel sends truncate before first close - FreeBSD only)
• (all) spelling corrections

• (mount) fixed rare race condition in read (not dangerous)
• (mount) better read/write synchronization in read
• (master) new chart (used/total space)