GlusterFS 11.0 veröffentlicht

GlusterFS ist eine Open-Source Software zur Erstellung und Konfiguration eines verteilten / distributed Dateisystems. Die Client-Server Architektur besteht aus mehren Komponenten und kann auf handelsüblicher Hardware erstellt werden.

GlusterFS 11.0 Release Notes

Announcements

1. Releases that receive maintenance updates post release 11 is 10 (reference)
2. Release 11 will receive maintenance updates around the 15th of every three months, and the release 10 will recieve maintainance updates around 15th every six months.

Highlights

• Major performance impovement of ~36% with rmdir operations #3685
• Extension of ZFS support for snapshots #2855
• Qouta implimentation based on namespace #1750
• Major cleanups and readdir/readdirp improvements link1 link2

Bugs addressed

Bugs addressed since release-10 are listed below:

• #1831 virtual images in replicated volume are not healed
• #1459 gluster_shared_storage failed to automount on node reboot on rhel 8
• #1458 sharding: fanout mknod process into multi threads
• #1457 systemd unit files missing from Debian 8.1-1 package
• #1456 virtual images in replicated volume are not healed?
• #1455 docs.gluster.org is down
• #1454 Geo-replication gsyncd at 100% CPU
• #1453 Disperse shd heal activity should be observable
• #1452 all glusterfs-client crashed at the same time
• #1451 Add details on ssl_setup_connection_params to help troubleshooting
• #1450 Please consider repackaging/providing dependency package (stubs)
• #145 Snapshot: improvements
• #1000 [bug:1193929] GlusterFS can be improved
• #1002 [bug:1679998] GlusterFS can be improved
• #1060 [bug:789278] Issues reported by Coverity static analysis tool
• #1686 mount-shared-storage.sh issue with systemd-automount
• #1757 RFE: improve namespace support in glusterfs
• #1774 RFE: simple-quota implementation
• #2123 Implement an I/O framework
• #2164 gf_proc_dump_call_stack() is not printing ctime correctly
• #2308 Developer sessions for glusterfs
• #2469 Python syntax error in syncdutils.py
• #2483 Place holder issue for fixing potential bugs in protocol/client/src
• #2491 Add s390x support to community CI
• #2664 Test suite produce uncompressed logs
• #2717 GlusterFS doesn’t support O_PATH flag in open()
• #2735 Remove byte-order.h and use the normal byteorder functions
• #2793 cluster.rebal-throttle description doesn’t seem to match the code
• #2832 selinux: make it possible to persist ganesha_use_fusefs from one up
• #2846 glusterd log filled with error messages.
• #2903 geo-rep restarts because of ‚list‘ object has no attribute ‚join‘ e
• #2910 glusterd: volume start doesn’t fail with properly if the port range
• #2912 When glusterfs uses an untrusted domain name， it cannot update the
• #2913 gluster lib cannot be dlopened: /lib64/libtcmalloc.so.4: cannot all
• #2916 Replace SHA deprecated functions with newer ones
• #2936 Wrong value for inodeSize in Volume status xml output.
• #2939 Volume log commands ‚filename‘ and ‚locate‘ described in the man bu
• #2944 tests: valid ip to be used instead of localhost, 127.0.0.1 or loopb
• #2947 IO failure when shrinking distributed dispersed volume while perfor
• #2962 cli: volume create without disperse count fails with ip addresses
• #2963 Do not use an iobuf pool for the CLI
• #2964 Cleanup the stub pool
• #2967 Make relevant functions static
• #2971 core file from /tests/basic/fencing/afr-lock-heal-advanced.t
• #2973 Allocate socket ioq outside of the out_lock
• #2975 quick-read: remove unused ’sh-failed‘ dict_get() function
• #2986 AFR: reduce variable indirection
• #2989 GF_ASSERT_AND_GOTO_WITH_ERROR improvements
• #2997 HA status is in FAILOVER when configuring NFS ganesha with pacemake
• #2998 Remove old authentication schemes
• #3004 Use gf_strndup() instead of gf_strdup() when string length is known
• #3005 multiple files: improve gf_dirent_for_name() functionality
• #3012 Reduce the impact of Cloudsync on IO path
• #3054 Update the links in gluster test framework documentation
• #3066 Online upgrade – 9.x to 10.0
• #3071 Log spam with glusterfs 10.0
• #3076 __lease_ctx_set() is not checking correctly for the return code
• #3103 glusterfs snapd crashes when snapshot is de-activated.
• #3130 Reduce the number of include statements
• #3137 Cleanup common include files
• #3166 bug: Incorrect Mac OS version assertion in configure.
• #3182 Some stale posix locks appear after a client disconnection
• #3187 Locks xlator leaks fd’s in some cases
• #3191 dht: Fix double free issue in the cbk function dht_common_mark_mdsx
• #3194 Log message for RPC clients is misleading because it logs unsigned
• #3213 Configure geo rep SSH with AES128 encryption – to improve performance
• #3217 Disperse volume with more than 16 data bricks fails to mount
• #3228 event-epoll.c: table_idx will always be 0?
• #3229 handle ‚option remote-port‘ properly in client protocol.
• #3232 The config for creating /var/run/gluster on boot is missing for glu
• #3234 SSL certificate wrong default path
• #3236 nfs: Optimize ctxcount value to reduce memory consumption for nfs-s
• #3240 dht_revalidate_cbk() needs to trigger directory heal with root perm
• #3248 fault in gluster command completion
• #3262 dht: Synchronize layout_(ref|unref) during layout_(get|set) in dht
• #3264 Posix private struct – reduce size
• #3288 GFID split brain resolution using favourite-child-policy leads to E
• #3294 Remove dht_nonblocking_inodelk() – it’s not used
• #3304 CID 1476381: (OVERRUN) @ /rpc/rpc-lib/src/rpc-clnt.c: 249 in __save
• #3321 server: Optimize server_process_event_upcall code path during handl
• #3329 mounting with ipv6 hostname leads to failure
• #3334 Error messages and 20 seconds timeout when creating qcow2 file via
• #3347 Test Failure: tests/bitrot/bug-1207627-bitrot-scrub-status.t
• #3359 meta_lookup() and other small improvements to meta xlator
• #3373 DHT doesn’t implement seek fop and causes failures
• #3375 glusterd: After node reboot not able to start all bricks successful
• #3382 Dictionary: remove all hash related code
• #3394 autoupdate – Update a configure.ac to a newer Autoconf
• #3417 Crash due to unaligned memory access
• #3426 logging.c small improvements
• #3469 Improve regression testing
• #3470 Spurious crash when „peer probing“ a non existing host name
• #3507 thin-arbiter-volume.t execute failed
• #3521 changelog: A brick process is getting crash due to SIGSEGV
• #3527 Brick process crashed when global thread pool is enabled
• #3604 is_nfs_export_available check and mount_nfs commands fail in RHEL 8
• #3636 posix: small storage environment and storage.reserve value
• #3647 wrong debug log in upcall_cache_invalidate()
• #3662 Some times Gluster Volume info XML returns wrong distCount
• #3683 dht: Cleanup linkto file by rebalance daemon while (hashed|cached)
• #3685 dht: Introduce rmdir-optimize option
• #3688 Infinite loop in dht when lookup fails with ENODATA
• #3695 test: ./tests/bugs/posix/bug-1651445.t is continuous failing during
• #3708 ./tests/basic/afr/afr-no-fsync.t is failing on FSYNCDIR, not on FSY
• #3710 Brick crashes automatically when writing small files
• #3717 syncop functions: no need to copy iatt structs if you are not going
• #3729 gf_svc_releasedir() logic error
• #3750 Bash completion is loaded every time bash starts
• #3774 Dead Lock In Brick Process
• #3778 test: Increase timeout for ./tests/basic/distribute/spare_file_reba
• #3781 Prime time client-only access control
• #3793 tests/bugs/replicate/bug-1586020-mark-dirty-for-entry-txn-on-quorum
• #3797 cdc xlator improvements
• #3823 rfc.sh: cannot detect upstream remote for non english locale
• #3831 afr: posix lock behavior is not correct while received an interrupt
• #3836 tests/bugs/glusterd/daemon-log-level-option.t fails on regression
• #3845 fuse_readdirp_cbk() – minor modifications
• #3847 gftest build failure
• #3855 reduce work for memory account
• #3876 fsetxattr() failed with EBADFD on opened directories
• #3891 Possible trash can size overflow

Quelle: https://docs.gluster.org/en/latest/release-notes/11.0/#bugs-addressed