RHEL8: samba server (smbd) repeatedly coredumps in open_directory ... TALLOC_FREE(lck) due to use after free or memory corruption
Issue
- On cifs client, a 'df' fails with 'Resource temporarily unavailable', 'Host is down' or 'Input/output error'
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
df: /mnt/client: Resource temporarily unavailable
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
df: /mnt/client: Host is down
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
df: /mnt/client: Input/output error
- Checking on the server side at the time of the cifs client error(s), we see the samba server (smbd) repeately coredumps
Environment
- Red Hat Enterprise Linux 8 (server)
- samba-4.10.4-101.el8_1.x86_64
- Red Hat Enterprise Linux 8 (client)
- Seen with kernel-4.18.0-147.4.1.el8_1
- share mounted without 'nohandlecache'
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.