Description of problem: After install node, then check the selinux log : there will be some AVC message in the log file. type=SYSCALL msg=audit(1359701540.208:211863): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fffdc48a8a0 a2=400 a3=0 items=0 ppid=1 pid=20921 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=5 comm="cgrulesengd" exe="/sbin/cgrulesengd" subj=unconfined_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359701540.208:211864): avc: denied { read } for pid=20921 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=unconfined_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir Version-Release number of selected component (if applicable): http://download.englab.nay.redhat.com/pub/rhel/rel-eng/RHEL6.4-20130130.0/6.4/Server/x86_64/ http://buildvm-devops.usersys.redhat.com/puddle/build/OpenShiftEnterprise/1.1.z/2013-01-30.2/ libselinux-utils-2.0.94-5.3.el6.x86_64 selinux-policy-3.7.19-195.el6.noarch selinux-policy-targeted-3.7.19-195.el6.noarch libselinux-2.0.94-5.3.el6.x86_64 How reproducible: always Steps to Reproduce: 1. Finish installing node 2. Check the log file "/var/log/audit/audit.log" Actual restuls: type=AVC msg=audit(1359701540.208:211864): avc: denied { read } for pid=20921 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=unconfined_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir Expected results: No such error Additional info: There will be more log in /var/log/message Feb 1 15:05:30 node CGRE[20921]: Warning: cannot write tid 31998 to /cgroup/all/openshift/46ad84fd669f464c8dc15e8e217abb47//tasks:No such process Feb 1 15:05:30 node CGRE[20921]: Warning: cgroup_attach_task_pid failed: 50016 Feb 1 15:05:30 node CGRE[20921]: Warning: failed to apply the rule. Error was: 50016 [root@node1 run]# ll -Z cgred.* -rw-r--r--. root root system_u:object_r:initrc_var_run_t:s0 cgred.pid srw-rw----. root cgred system_u:object_r:cgred_var_run_t:s0 cgred.socket
cat'ing those two lines into "audit2allow -r" suggests the following. I don't think this is OpenShift related so I'm going to search BZ for an upstream bug. require { type cgred_t; type inotifyfs_t; class dir read; } #============= cgred_t ============== allow cgred_t inotifyfs_t:dir read;
Are you getting only these AVC msgs? Could you execute # setenforce 0 re-test it # ps -eZ |grep initrc # ausearch -m avc -ts recent # setenforce 1
[root@node ~]# setenforce 0 [root@node ~]# ps -eZ |grep initrc system_u:system_r:openshift_initrc_t:s0-s0:c0.c1023 1496 ? 00:06:12 ruby system_u:system_r:openshift_initrc_t:s0-s0:c0.c1023 15360 ? 00:00:00 sh <defunct> system_u:system_r:openshift_initrc_t:s0-s0:c0.c1023 21723 ? 00:00:00 haproxy [root@node ~]# ausearch -m avc -ts recent ---- time->Mon Feb 4 06:47:06 2013 type=SYSCALL msg=audit(1359978426.667:16668336): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fff71f2c000 a2=400 a3=0 items=0 ppid=1 pid=1367 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cgrulesengd" exe=2F7362696E2F636772756C6573656E6764202864656C6574656429 subj=system_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359978426.667:16668336): avc: denied { read } for pid=1367 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=system_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir ---- time->Mon Feb 4 06:47:06 2013 type=SYSCALL msg=audit(1359978426.667:16668337): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fff71f2c000 a2=400 a3=0 items=0 ppid=1 pid=1367 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cgrulesengd" exe=2F7362696E2F636772756C6573656E6764202864656C6574656429 subj=system_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359978426.667:16668337): avc: denied { read } for pid=1367 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=system_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir ---- time->Mon Feb 4 06:47:06 2013 type=SYSCALL msg=audit(1359978426.667:16668338): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fff71f2c000 a2=400 a3=0 items=0 ppid=1 pid=1367 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cgrulesengd" exe=2F7362696E2F636772756C6573656E6764202864656C6574656429 subj=system_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359978426.667:16668338): avc: denied { read } for pid=1367 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=system_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir ---- time->Mon Feb 4 06:47:06 2013 type=SYSCALL msg=audit(1359978426.667:16668339): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fff71f2c000 a2=400 a3=0 items=0 ppid=1 pid=1367 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cgrulesengd" exe=2F7362696E2F636772756C6573656E6764202864656C6574656429 subj=system_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359978426.667:16668339): avc: denied { read } for pid=1367 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=system_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir ---- time->Mon Feb 4 06:47:06 2013 type=SYSCALL msg=audit(1359978426.667:16668340): arch=c000003e syscall=0 success=no exit=-13 a0=3 a1=7fff71f2c000 a2=400 a3=0 items=0 ppid=1 pid=1367 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cgrulesengd" exe=2F7362696E2F636772756C6573656E6764202864656C6574656429 subj=system_u:system_r:cgred_t:s0 key=(null) type=AVC msg=audit(1359978426.667:16668340): avc: denied { read } for pid=1367 comm="cgrulesengd" path="inotify" dev=inotifyfs ino=1 scontext=system_u:system_r:cgred_t:s0 tcontext=system_u:object_r:inotifyfs_t:s0 tclass=dir <--snip-->
Does everything work correctly with a local policy?
Hi Miroslav, What do you mean by a local policy exactly?
I meant # grep cgruleseng /var/log/audit/audit.log |audit2allow -R -M mypol # semodule -i mypol.pp
Yes, after running those commands the denials stopped happening. Here is the module that was generated: policy_module(mypol, 1.0) require { type cgred_t; } #============= cgred_t ============== fs_list_inotifyfs(cgred_t)
ok, we need to create a policy bug.
I'm moving this out of 1.1.1 since it will be fixed in base RHEL. I have cloned the bug there.
Version: http://buildvm-devops.usersys.redhat.com/puddle/build/OpenShiftEnterprise/1.1.z/2013-03-06.1/ Verify： 1) verifying cgrulesengd denials and 100% cpu utilization on my dev environment 2) upgrade to selinux-policy-mls-3.7.19-195.el6_4.2.noarch.rpm [root@node ~]# rpm -qa | grep selinux | grep policy selinux-policy-targeted-3.7.19-195.el6_4.2.noarch [root@node ~]# grep SELINUXTYPE /etc/selinux/config # SELINUXTYPE= can take one of these two values: SELINUXTYPE=targeted 3) smoke tested that everything worked as expected without having to restart cgrulesengd https://tcms.engineering.redhat.com/run/58770/ I will create another bug for tracing the warning message in /var/log/message.
I'm closing this bug since the actual fix shipped with RHEL6.4. Thanks for the help!