We detected a valgrind error as shown below, which should be fixed to potentially cause rare db connection issues (note the line numbers, based on source code appear slightly off):

==13975== Invalid read of size 1
==13975== at 0x12B75867: client_mpvio_read_packet (my_auth.c:474)
==13975== by 0x12B74B96: native_password_auth_client (my_auth.c:117)
==13975== by 0x12B74DB3: run_plugin_auth (my_auth.c:646)
==13975== by 0x12B5F278: mthd_my_real_connect (libmariadb.c:2042)
...

==13975== Address 0x14b26250 is 0 bytes inside a block of size 8,192 free'd
==13975== at 0x4A06430: free (vg_replace_malloc.c:446)
==13975== by 0x12B58F5C: net_end (net.c:144)
==13975== by 0x12B5CCFF: end_server (libmariadb.c:812)
==13975== by 0x12B5DE8C: net_safe_read (libmariadb.c:414)
==13975== by 0x12B75857: client_mpvio_read_packet (my_auth.c:469)
==13975== by 0x12B74B96: native_password_auth_client (my_auth.c:117)
==13975== by 0x12B74DB3: run_plugin_auth (my_auth.c:646)
==13975== by 0x12B5F278: mthd_my_real_connect (libmariadb.c:2042)