[LU-6895] sanity-lfsck test 4 hung: bad entry in directory: rec_len is smaller than minimal - inode=3925999616

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.8.0
• Affects Version/s: Lustre 2.8.0
• Environment: Lustre Tag: 2.7.56; Distro/Arch: SLES11SP3/x86_64 (both client and server)
• Severity: 3
• Assignee: nasf (Inactive)
• Reporter: Jian Yu

Description

sanity-lfsck test 4 hung as follows:

Starting client: shadow-23vm5:  -o user_xattr,flock shadow-23vm8@tcp:/lustre /mnt/lustre
CMD: shadow-23vm5 mkdir -p /mnt/lustre
CMD: shadow-23vm5 mount -t lustre -o user_xattr,flock shadow-23vm8@tcp:/lustre /mnt/lustre

On client shadow-23vm5:

[ 9794.882548] Lustre: DEBUG MARKER: mkdir -p /mnt/lustre
[ 9794.888889] Lustre: DEBUG MARKER: mount -t lustre -o user_xattr,flock shadow-23vm8@tcp:/lustre /mnt/lustre
[ 9794.910641] LustreError: 11-0: lustre-MDT0000-mdc-ffff880071b43400: operation mds_connect to node 10.1.5.29@tcp failed: rc = -30
[ 9799.909475] LustreError: 11-0: lustre-MDT0000-mdc-ffff880071b43400: operation mds_connect to node 10.1.5.29@tcp failed: rc = -30

On MDS shadow-23vm8:

[ 9799.279381] Lustre: DEBUG MARKER: /usr/sbin/lctl get_param -n mdd.lustre-MDT0000.lfsck_namespace
[ 9799.343571] LustreError: 9974:0:(tgt_lastrcvd.c:929:tgt_client_new()) lustre-MDT0000: unable to set MULTI RPCS incompatibility flag
[ 9801.154230] LustreError: 10003:0:(qsd_entry.c:291:qsd_update_index()) lustre-MDT0000: failed to update global index copy for id 0, rc:-30
[ 9801.154261] Lustre: 10003:0:(qsd_reint.c:487:qsd_reint_main()) lustre-MDT0000: reint global for [0x200000006:0x1010000:0x0] failed. -30
[ 9804.341804] LustreError: 9974:0:(tgt_lastrcvd.c:952:tgt_client_new()) lustre-MDT0000: Failed to write client lcd at idx 0, rc -30
[ 9809.342270] LustreError: 9974:0:(tgt_lastrcvd.c:952:tgt_client_new()) lustre-MDT0000: Failed to write client lcd at idx 0, rc -30
[ 9814.342330] LustreError: 9974:0:(tgt_lastrcvd.c:952:tgt_client_new()) lustre-MDT0000: Failed to write client lcd at idx 0, rc -30

Maloo report: https://testing.hpdd.intel.com/test_sets/2642815a-2f37-11e5-bc70-5254006e85c2

Issue Links

• is blocking: LU-6843 Support RHEL7 on b2_5_fe
• is duplicated by: LU-7572 sanity-scrub 4c : server kernel BUG at lustre-2.7.64/ldiskfs/namei.c:208
• is related to: LU-7169 conf-sanity 84 restart mds1 failed
• is related to: LU-7315 In osd_dirent_check_repair(), Pointer 'hlock' can be NULL and can be dereferenced