ceph-crash: drop privleges to run as "ceph" user, rather than root

If privileges cannot be dropped, log an error and exit.  This commit
also catches and logs exceptions when scraping the crash path, without
which ceph-crash would just exit if it encountered an error.

Fixes: CVE-2022-3650
Fixes: https://tracker.ceph.com/issues/57967
Signed-off-by: Tim Serong <tserong@suse.com>

src/ceph-crash.in

@@ -3,8 +3,10 @@
 # vim: ts=4 sw=4 smarttab expandtab
 
 import argparse
+import grp
 import logging
 import os
+import pwd
 import signal
 import socket
 import subprocess
@@ -88,8 +90,25 @@ def handler(signum, frame):
     sys.exit(0)
 
 
+def drop_privs():
+    if os.getuid() == 0:
+        try:
+            ceph_uid = pwd.getpwnam("ceph").pw_uid
+            ceph_gid = grp.getgrnam("ceph").gr_gid
+            os.setgroups([])
+            os.setgid(ceph_gid)
+            os.setuid(ceph_uid)
+        except Exception as e:
+            log.error(f"Unable to drop privileges: {e}")
+            sys.exit(1)
+
+
 def main():
     global auth_names
+
+    # run as unprivileged ceph user
+    drop_privs()
+
     # exit code 0 on SIGINT, SIGTERM
     signal.signal(signal.SIGINT, handler)
     signal.signal(signal.SIGTERM, handler)
@@ -112,7 +131,10 @@ def main():
 
     log.info("monitoring path %s, delay %ds" % (args.path, args.delay * 60.0))
     while True:
-        scrape_path(args.path)
+        try:
+            scrape_path(args.path)
+        except Exception as e:
+            log.error(f"Error scraping {args.path}: {e}")
         if args.delay == 0:
             sys.exit(0)
         time.sleep(args.delay * 60)