Bug #36522: ceph-mon memory leak from dpdk - Ceph - Ceph

Actions

Copy link

Bug #36522

closed

ceph-mon memory leak from dpdk

Added by Casey Bodley over 5 years ago. Updated over 5 years ago.

Status:

Resolved

Priority:

Normal

Assignee:

Category:

Target version:

% Done:

Source:

Tags:

ceph-mon valgrind

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

the rgw suite is seeing some valgrind issues in ceph-mon related to rte_log_register(). each stack is ~~slightly~~ different, but different enough that it doesn't match the suppression added in https://github.com/ceph/teuthology/pull/1220.

<error>
  <unique>0x1</unique>
  <tid>1</tid>
  <kind>Leak_StillReachable</kind>
  <xwhat>
    <text>11 bytes in 1 blocks are still reachable in loss record 2 of 30</text>
    <leakedbytes>11</leakedbytes>
    <leakedblocks>1</leakedblocks>
  </xwhat>
  <stack>
    <frame>
      <ip>0x4C29C23</ip>
      <obj>/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so</obj>
      <fn>malloc</fn>
      <dir>/builddir/build/BUILD/valgrind-3.13.0/coregrind/m_replacemalloc</dir>
      <file>vg_replace_malloc.c</file>
      <line>299</line>
    </frame>
    <frame>
      <ip>0x115F3809</ip>
      <obj>/usr/lib64/libc-2.17.so</obj>
      <fn>strdup</fn>
    </frame>
    <frame>
      <ip>0x60D8D2</ip>
      <obj>/usr/bin/ceph-mon</obj>
      <fn>rte_log_register</fn>
    </frame>
    <frame>
      <ip>0x2B087F</ip>
      <obj>/usr/bin/ceph-mon</obj>
    </frame>
    <frame>
      <ip>0x870B2C</ip>
      <obj>/usr/bin/ceph-mon</obj>
      <fn>__libc_csu_init</fn>
    </frame>
    <frame>
      <ip>0x115893D4</ip>
      <obj>/usr/lib64/libc-2.17.so</obj>
      <fn>(below main)</fn>
    </frame>
  </stack>
</error>

<error>
  <unique>0x14</unique>
  <tid>1</tid>
  <kind>Leak_StillReachable</kind>
  <xwhat>
    <text>249 bytes in 29 blocks are still reachable in loss record 21 of 30</text>
    <leakedbytes>249</leakedbytes>
    <leakedblocks>29</leakedblocks>
  </xwhat>
  <stack>
    <frame>
      <ip>0x4C29C23</ip>
      <obj>/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so</obj>
      <fn>malloc</fn>
      <dir>/builddir/build/BUILD/valgrind-3.13.0/coregrind/m_replacemalloc</dir>
      <file>vg_replace_malloc.c</file>
      <line>299</line>
    </frame>
    <frame>
      <ip>0x115F3809</ip>
      <obj>/usr/lib64/libc-2.17.so</obj>
      <fn>strdup</fn>
    </frame>
    <frame>
      <ip>0x2B07E4</ip>
      <obj>/usr/bin/ceph-mon</obj>
    </frame>
    <frame>
      <ip>0x870B2C</ip>
      <obj>/usr/bin/ceph-mon</obj>
      <fn>__libc_csu_init</fn>
    </frame>
    <frame>
      <ip>0x115893D4</ip>
      <obj>/usr/lib64/libc-2.17.so</obj>
      <fn>(below main)</fn>
    </frame>
  </stack>
</error>

<error>
  <unique>0x17</unique>
  <tid>1</tid>
  <kind>Leak_StillReachable</kind>
  <xwhat>
    <text>528 bytes in 1 blocks are still reachable in loss record 24 of 30</text>
    <leakedbytes>528</leakedbytes>
    <leakedblocks>1</leakedblocks>
  </xwhat>
  <stack>
    <frame>
      <ip>0x4C2BBB8</ip>
      <obj>/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so</obj>
      <fn>realloc</fn>
      <dir>/builddir/build/BUILD/valgrind-3.13.0/coregrind/m_replacemalloc</dir>
      <file>vg_replace_malloc.c</file>
      <line>785</line>
    </frame>
    <frame>
      <ip>0x60D8BB</ip>
      <obj>/usr/bin/ceph-mon</obj>
      <fn>rte_log_register</fn>
    </frame>
    <frame>
      <ip>0x2B087F</ip>
      <obj>/usr/bin/ceph-mon</obj>
    </frame>
    <frame>
      <ip>0x870B2C</ip>
      <obj>/usr/bin/ceph-mon</obj>
      <fn>__libc_csu_init</fn>
    </frame>
    <frame>
      <ip>0x115893D4</ip>
      <obj>/usr/lib64/libc-2.17.so</obj>
      <fn>(below main)</fn>
    </frame>
  </stack>
</error>

example: http://qa-proxy.ceph.com/teuthology/cbodley-2018-10-16_17:32:21-rgw-wip-rgw-putobjprocessor-distro-basic-smithi/3148539/remote/smithi184/log/valgrind/c1.mon.a.log.gz

Related issues 1 (0 open — 1 closed)