[LU-11627] LustreError: 86028:0:(file.c:454:ll_dom_finish_open()) ASSERTION( lnb.lnb_file_offset % (1UL << 16) == 0 ) failed:

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.12.0
• Type: Bug
• Priority: Major
• Reporter: James A Simmons
• Assignee: Jian Yu
• Resolution: Duplicate

Environment

ARM/Power8 clients running latest pre-2.12 lustre

• Severity: 3

Description

While testing ARM/Power8 I saw the following crash in sanity 271f:

[Mon Nov  5 23:24:34 2018][13642.788756] LustreError: 86028:0:(file.c:454:ll_dom_finish_open()) ASSERTION( lnb.lnb_file_offset % (1UL << 16) == 0 ) failed:

[Mon Nov  5 23:24:34 2018][13642.788923] LustreError: 86028:0:(file.c:454:ll_dom_finish_open()) LBUG

[Mon Nov  5 23:24:34 2018][13642.788983] Pid: 86028, comm: cat 4.14.0-49.6.1.el7a.ppc64le #1 SMP Wed May 16 21:05:05 UTC 2018

[Mon Nov  5 23:24:34 2018][13642.789065] Call Trace:

[Mon Nov  5 23:24:34 2018][13642.789101]  libcfs_call_trace+0x98/0xf0 [libcfs]

[Mon Nov  5 23:24:34 2018][13642.789154]  lbug_with_loc+0x5c/0xc0 [libcfs]

[Mon Nov  5 23:24:34 2018][13642.789215]  ll_dom_finish_open+0xa48/0xb40 [lustre]

[Mon Nov  5 23:24:34 2018][13642.789273]  ll_lookup_it_finish+0x6d0/0xe40 [lustre]

[Mon Nov  5 23:24:34 2018][13642.789331]  ll_lookup_it+0x4ac/0xf10 [lustre]

[Mon Nov  5 23:24:34 2018][13642.789389]  ll_atomic_open+0x264/0xdb0 [lustre]

[Mon Nov  5 23:24:34 2018][13642.789439]  lookup_open+0x200/0x780

[Mon Nov  5 23:24:34 2018][13642.789475]  path_openat+0x824/0x1070

[Mon Nov  5 23:24:34 2018][13642.789512]  do_filp_open+0x88/0x140

[Mon Nov  5 23:24:34 2018][13642.789549]  SyS_open+0x1bc/0x300

[Mon Nov  5 23:24:34 2018][13642.789586]  system_call+0x58/0x6c

[Mon Nov  5 23:24:34 2018][13642.789622] Kernel panic - not syncing: LBUG

[Mon Nov  5 23:24:34 2018][13642.789671] CPU: 97 PID: 86028 Comm: cat Tainted: G           OE  ------------   4.14.0-49.6.1.el7a.ppc64le #1

[Mon Nov  5 23:24:34 2018][13642.789765] Call Trace:

[Mon Nov  5 23:24:34 2018][13642.789792] [c0000017e6773530] [c000000000c3fdcc] dump_stack+0xb0/0xf4 (unreliable)

[Mon Nov  5 23:24:34 2018][13642.789864] [c0000017e6773570] [c000000000136d64] panic+0x150/0x32c

[Mon Nov  5 23:24:34 2018][13642.789929] [c0000017e6773600] [d000000013090d38] lbug_with_loc+0xb8/0xc0 [libcfs]

[Mon Nov  5 23:24:34 2018][13642.790010] [c0000017e6773670] [d0000000187d0558] ll_dom_finish_open+0xa48/0xb40 [lustre]

[Mon Nov  5 23:24:34 2018][13642.790091] [c0000017e6773780] [d0000000188145f0] ll_lookup_it_finish+0x6d0/0xe40 [lustre]

Issue Links

• duplicates: LU-11595 sanity-dom sanityn test 11: LBUG: (fi...
• duplicates: LU-12014 check correct size in ll_dom_finish_o...

Comments

[Peter Jones]

Jian

Could you please advise?

Thanks

Peter

[Jian Yu]

Sure, Peter.

The failure is related to large page size of ARM processor:

# getconf PAGE_SIZE
65536

And the codes that caused the assertion are:

ll_dom_finish_open()

        lnb.lnb_file_offset = rnb->rnb_offset;
        start = lnb.lnb_file_offset / PAGE_SIZE;
        index = 0;
        LASSERT(lnb.lnb_file_offset % PAGE_SIZE == 0);

Let me close this ticket as a duplicate of LU-11595 so that Mike can advise how to fix the above DOM codes.