I'm intern from SElinux team at Red Hat and I write new SElinux policy for your service, because the service doesnt have policy.
Here  I send how look SElinux denial:

type=AVC msg=audit(1565874853.606:832): avc:  denied  { sys_admin } for  pid=9046 comm="stratisd" capability=21  scontext=system_u:system_r:stratisd_t:s0 tcontext=system_u:system_r:stratisd_t:s0 tclass=capability permissive=0

It means that process stratisd require sysadmin capability.

man capabilities:
For  the  purpose of performing permission checks, traditional UNIX implementations dis‐
tinguish two categories of processes: privileged processes (whose effective user  ID  is
0, referred to as superuser or root), and unprivileged processes (whose effective UID is
nonzero).  Privileged processes bypass all kernel permission checks, while  unprivileged
processes  are  subject  to  full permission checking based on the process's credentials
(usually: effective UID, effective GID, and supplementary group list).

Sysadmin capability:

             * Perform  a  range  of  system administration operations including: quotactl(2),
                mount(2), umount(2),  swapon(2),  swapoff(2),  sethostname(2),  and  setdomain‐
              * perform  privileged syslog(2) operations (since Linux 2.6.37, CAP_SYSLOG should
                be used to permit such operations);
              * perform VM86_REQUEST_IRQ vm86(2) command;
              * perform IPC_SET and IPC_RMID operations on arbitrary System V IPC objects;
              * override RLIMIT_NPROC resource limit;
              * perform operations on trusted and security Extended Attributes (see xattr(7));
              * use lookup_dcookie(2);
              * use  ioprio_set(2)  to  assign  IOPRIO_CLASS_RT  and  (before   Linux   2.6.25)
                IOPRIO_CLASS_IDLE I/O scheduling classes;
              * forge PID when passing socket credentials via UNIX domain sockets;
              * exceed  /proc/sys/fs/file-max,  the  system-wide  limit  on  the number of open
                files, in system calls that open files (e.g.,  accept(2),  execve(2),  open(2),
              * employ  CLONE_*  flags  that create new namespaces with clone(2) and unshare(2)
                (but, since Linux 3.8, creating user namespaces does not require  any  capabil‐
              * call perf_event_open(2);
              * access privileged perf event information;
              * call setns(2) (requires CAP_SYS_ADMIN in the target namespace);
              * call fanotify_init(2);
              * call bpf(2);
              * perform privileged KEYCTL_CHOWN and KEYCTL_SETPERM keyctl(2) operations;
              * perform madvise(2) MADV_HWPOISON operation;
              * employ the TIOCSTI ioctl(2) to insert characters into the input queue of a ter‐
                minal other than the caller's controlling terminal;
              * employ the obsolete nfsservctl(2) system call;
              * employ the obsolete bdflush(2) system call;
              * perform various privileged block-device ioctl(2) operations;
              * perform various privileged filesystem ioctl(2) operations;
              * perform privileged ioctl(2) operations on  the  /dev/random  device  (see  ran‐
              * install a seccomp(2) filter without first having to set the no_new_privs thread
              * modify allow/deny rules for device control groups;
              * employ the ptrace(2) PTRACE_SECCOMP_GET_FILTER operation to dump             tracee's  seccomp filters;
              * employ  the  ptrace(2) PTRACE_SETOPTIONS operation to suspend the tracee's     seccomp protections (i.e., the PTRACE_O_SUSPEND_SECCOMP flag);
              * perform administrative operations on many device drivers.

Need stratisd service one of these operations?


Patrik Koncity

Intern, Security Technologies

Red Hat, Inc.