Lustre 2.10.7 Changelog: Difference between revisions

From Lustre Wiki
Jump to navigation Jump to search
mNo edit summary
m (→‎Issues fixed between 2.10.6 and 2.10.7: list new features and improvements before bug fixes)
 
Line 15: Line 15:


== Issues fixed between 2.10.6 and 2.10.7 ==
== Issues fixed between 2.10.6 and 2.10.7 ==
''' Improvement '''
* [[https://jira.whamcloud.com/browse/LU-8402 LU-8402]] - OBD_CONNECT_* flags retirement
* [[https://jira.whamcloud.com/browse/LU-8727 LU-8727]] - Remove skip records from config file
* [[https://jira.whamcloud.com/browse/LU-9810 LU-9810]] - Melanox OFED 4.1 support
* [[https://jira.whamcloud.com/browse/LU-10507 LU-10507]] - Use {save,restore}_layout() to save and restore layout in test cases
* [[https://jira.whamcloud.com/browse/LU-10516 LU-10516]] - Update recommended e2fsprogs version
* [[https://jira.whamcloud.com/browse/LU-11010 LU-11010]] - Remove call to ‘return’ after call to ‘skip’ in Lustre test suites
* [[https://jira.whamcloud.com/browse/LU-11244 LU-11244]] - 'make debs' mistakenly ignores IB options
* [[https://jira.whamcloud.com/browse/LU-11736 LU-11736]] - Do not apply bulk IO tuning on MDT or MGT devices
* [[https://jira.whamcloud.com/browse/LU-11739 LU-11739]] - Don't inherit default layout from root for new subdirectories


''' Bug '''
''' Bug '''
Line 82: Line 93:
''' Task '''
''' Task '''
* [[https://jira.whamcloud.com/browse/LU-11689 LU-11689]] - lfs project should proceed all dirs if hit errors
* [[https://jira.whamcloud.com/browse/LU-11689 LU-11689]] - lfs project should proceed all dirs if hit errors
''' Improvement '''
* [[https://jira.whamcloud.com/browse/LU-8402 LU-8402]] - OBD_CONNECT_* flags retirement
* [[https://jira.whamcloud.com/browse/LU-8727 LU-8727]] - Remove skip records from config file
* [[https://jira.whamcloud.com/browse/LU-9810 LU-9810]] - Melanox OFED 4.1 support
* [[https://jira.whamcloud.com/browse/LU-10507 LU-10507]] - Use {save,restore}_layout() to save and restore layout in test cases
* [[https://jira.whamcloud.com/browse/LU-10516 LU-10516]] - Update recommended e2fsprogs version
* [[https://jira.whamcloud.com/browse/LU-11010 LU-11010]] - Remove call to ‘return’ after call to ‘skip’ in Lustre test suites
* [[https://jira.whamcloud.com/browse/LU-11244 LU-11244]] - 'make debs' mistakenly ignores IB options
* [[https://jira.whamcloud.com/browse/LU-11736 LU-11736]] - Do not apply bulk IO tuning on MDT or MGT devices
* [[https://jira.whamcloud.com/browse/LU-11739 LU-11739]] - Don't inherit default layout from root for new subdirectories

Latest revision as of 16:11, 18 November 2019

Server support for kernels:

  • RHEL 7.6 (3.10.0-957.1.3.el7)

Client support for unpatched kernels:

  • RHEL 6.10 (2.6.32-754.10.1.el6)
  • RHEL 7.6 (3.10.0-957.1.3.el7)
  • SLES12 SP3 (4.4.162-94.69)
  • Ubuntu 16.04 (4.4.0-142)

Recommended e2fsprogs version:

  • v1.44.5.wc1

Interoperability Support:

  • Clients & Servers: Latest 2.9.X

Issues fixed between 2.10.6 and 2.10.7

Improvement

  • [LU-8402] - OBD_CONNECT_* flags retirement
  • [LU-8727] - Remove skip records from config file
  • [LU-9810] - Melanox OFED 4.1 support
  • [LU-10507] - Use {save,restore}_layout() to save and restore layout in test cases
  • [LU-10516] - Update recommended e2fsprogs version
  • [LU-11010] - Remove call to ‘return’ after call to ‘skip’ in Lustre test suites
  • [LU-11244] - 'make debs' mistakenly ignores IB options
  • [LU-11736] - Do not apply bulk IO tuning on MDT or MGT devices
  • [LU-11739] - Don't inherit default layout from root for new subdirectories

Bug

  • [LU-4939] - Need to be able to sanely query and change MGS configuration information
  • [LU-6632] - lctl replace_nids should remove SKIP records, not EXCLUDE ones
  • [LU-7558] - niobuf.c:721:ptl_send_rpc() LASSERT(AT_OFF || imp_state != LUSTRE_IMP_FULL || imp_msghdr_flags & MSGHDR_AT_SUPPORT ...)
  • [LU-7631] - conf-sanity test_82a: getstripe -c wrong: found 2, expected 3
  • [LU-7988] - HSM: high lock contention for cdt_llog_lock
  • [LU-8895] - server grants clients with more grants that the clients ask
  • [LU-8999] - sanity-quota test_38: skipped id entries
  • [LU-9230] - soft lockup on v2.9 Lustre clients (ldlm?)
  • [LU-9474] - tests: register traps in a correct order
  • [LU-9706] - conf-sanity test_53a: MDS soft lockup - CPU#0 stuck for 22s in osd_trans_commit_cb()
  • [LU-9785] - lov_io_slice_init()) ASSERTION( obj->lo_lsm != ((void *)0) ) failed
  • [LU-9874] - simplify project transfer codes
  • [LU-9906] - Allow Lustre page dropping to use pagevec_release
  • [LU-10055] - mds_fill_lvbo() message spew on MDS console
  • [LU-10124] - lnetctl: lnetctl import --add not importing peers corectly
  • [LU-10143] - LBUG dt_object.h:2166:dt_declare_record_write
  • [LU-10213] - o2iblnd: Potential discrepancy when allocating qp
  • [LU-10461] - sanity 77c failure due to variables called in cleanup routine are out of scope
  • [LU-10527] - LustreError: 7830:0:(llog_cat.c:313:llog_cat_current_log()) ASSERTION( llh )
  • [LU-10565] - Handle some small features in upstream
  • [LU-10612] - replay-single test_48 defect causes test failure due to error: No space left on device
  • [LU-10629] - Cannot clear a default OST pool from root
  • [LU-10776] - Large grant requests still don't work, resulting in small write RPCs
  • [LU-10818] - mds-survey test 2 hangs with “ASSERTION( ma->ma_need & (MA_LOV | MA_LMV) ) failed”
  • [LU-10900] - wrong assertion in osd_transfer_project()
  • [LU-10945] - Race between sending bl ast and lock cancel
  • [LU-10985] - Attempting to send a mkdir create intents crashes server
  • [LU-10986] - make lfs project tolerant errors
  • [LU-11006] - lnetctl export creates "peer" root with no peer
  • [LU-11049] - lgss-sk: incorrect 'fsname' and 'nodemap' string handling
  • [LU-11056] - OSS can't connect to MDS after hard reboot
  • [LU-11068] - "not a valid identifier" in lbuild caused by an extra space
  • [LU-11130] - cross-target rename creates invalid symlink inodes
  • [LU-11156] - scrub treat project quota inode as IGIF last_id
  • [LU-11158] - PFL component instantiation is not replayed properly
  • [LU-11186] - Cannot create files on MDT1, ofd_precreate_objects()) snx11205-OST0001:0xffffffdd:11811161088 hit the OBIF_MAX_OID (1<<32)!
  • [LU-11187] - MMP updated sometimes failes T10PI checks
  • [LU-11195] - LBUG: (lod_lov.c:633:lod_free_comp_buffer()) ASSERTION( entry->llc_stripe == ((void *)0) )
  • [LU-11196] - conf-sanity test_103: Fail to cleanup the env
  • [LU-11461] - Lustre resource agent fails with symlink target
  • [LU-11481] - corrupt directory
  • [LU-11536] - ofd_create_hdl returns 0 instead of ENOSPC causing ESTALE on MDS
  • [LU-11567] - llog_reader does not show index of changelog records
  • [LU-11579] - cl_file_inode_init()) ASSERTION(inode->i_state & (1 << 3) ) failed:
  • [LU-11582] - LBUG: ASSERTION( inode->i_data.nrpages == 0 ) failed
  • [LU-11605] - create_count stuck in 0 after changeing max_create_count to 0 and back 20 000
  • [LU-11620] - BUG: sleeping function called from invalid context at mm/slub.c:940
  • [LU-11625] - ls -l returns "ls: cannot access Invalid argument"
  • [LU-11647] - niobuf.c:330:ptlrpc_register_bulk()) ASSERTION( desc->bd_md_count == 0 ) failed:
  • [LU-11652] - kernel update [SLES12 SP3 4.4.162-94.69.2]
  • [LU-11658] - The cl_cache may be uninitialized while osc is activating
  • [LU-11720] - SRPM unnecessarily tied to kernel version
  • [LU-11723] - kernel update [RHEL7.6 3.10.0-957.1.3.el7]
  • [LU-11737] - LustreError: 11060:0:(osd_handler.c:3985:osd_xattr_set()) ASSERTION( handle ) failed:
  • [LU-11757] - lfs mkdir: Cannot allocate memory
  • [LU-11783] - lustre/lustre_user.h is deprecated too soon
  • [LU-11790] - ext4: include terminating u32 in size of xattr entries when expanding inodes
  • [LU-11870] - kernel update [RHEL6.10 2.6.32-754.10.1.el6]
  • [LU-11943] - many input/output error after soak running for couple of hours
  • [LU-11947] - Regression caused by LU-11461
  • [LU-12065] - Client got evicted when lock callback timer expired on OSS
  • [LU-12073] - conf-sanity test 123aa hangs

Task

  • [LU-11689] - lfs project should proceed all dirs if hit errors