Bug 741455

Summary: inability to read/write a file should not print parse error, but instead print that the file couldn't be read/written
Product: Red Hat Enterprise Linux 6 Reporter: Steven Dake <sdake>
Component: corosyncAssignee: Jan Friesse <jfriesse>
Status: CLOSED ERRATA QA Contact: Cluster QE <mspqa-list>
Severity: low Docs Contact:
Priority: low    
Version: 6.2CC: ccaulfie, cluster-maint, cmarthal, dwalsh, jkortus, lhh, rpeterso, teigland
Target Milestone: rc   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: corosync-1.4.1-6.el6 Doc Type: Bug Fix
Doc Text:
Cause Set logging to file if path to file contains nonexisting directory. Consequence Error message which doesn't make sense (configuration file error) Fix Pass proper string pointer to function for opening log file Result Correct error message (can't create log file) is displayed.
 Story Points: ---
Clone Of: 741434 Environment:
Last Closed: 2012-06-20 12:22:37 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On: 741434    
Bug Blocks: 756082    
Attachments:
Description Flags
Proposed patch none

Description Steven Dake 2011-09-26 20:59:17 UTC 
+++ This bug was initially created as a clone of Bug #741434 +++

Description of problem:
Every once and awhile I'm seeing the cluster fail to start due to "parse errors", however if I run the actual cman_tool join cmd by itself, the cluster starts just fine. Also, I find it odd that there is a "parse error" right after seeing the "Successfully parsed cman config" message.

[root@taft-01 tmp]# service cman start
Starting cluster: 
   Checking if cluster has been disabled at boot...        [  OK  ]
   Checking Network Manager...                             [  OK  ]
   Global setup...                                         [  OK  ]
   Enable Xend bridge net workaround... action not required
   Loading kernel modules...                               [  OK  ]
   Mounting configfs...                                    [  OK  ]
   Starting cman... 
   corosync died: Could not read cluster configuration Check cluster logs for details
                                                           [FAILED]


Sep 26 14:44:17 taft-01 corosync[4787]: parse error in config: parse error in config: .
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] Corosync Cluster Engine ('1.4.1'): started and ready to provide service.
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] Corosync built-in features: nss dbus rdma snmp
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] Successfully read config from /etc/cluster/cluster.conf
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] Successfully parsed cman config
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] parse error in config: parse error in config: .
Sep 26 14:44:17 taft-01 corosync[4787]:   [MAIN  ] Corosync Cluster Engine exiting with status 8 at main.c:1680.


However if I run the actual cman_tool join cmd by itself, the cluster joins just fine.

# cman_tool -t 60 -w join  -DWARN


Version-Release number of selected component (if applicable):
Linux taft-01 2.6.32-198.el6.x86_64 #1 SMP Thu Sep 15 23:40:38 EDT 2011 x86_64 x86_64 x86_64 GNU/Linux
corosync-1.4.1-3.el6.x86_64

How reproducible:
service cman start

--- Additional comment from cmarthal on 2011-09-26 16:09:19 EDT ---

Created attachment 524979 [details]
failing cluster.conf file

--- Additional comment from sdake on 2011-09-26 16:54:14 EDT ---

type=AVC msg=audit(1316212684.082:172): avc:  denied  { read } for  pid=8091 comm="corosync" name="corosync.log" dev=dm-0 ino=131142 scontext=unconfined_u:system_r:corosync_t:s0 tcontext=system_u:object_r:var_log_t:s0 tclass=file

not a corosync problem.  Corosync doesn't modify selinux settings.  restorecon -R /var/log/messages fixed the problem.

-rw-r--r--. root root system_u:object_r:corosync_var_log_t:s0 corosync.log

The audit log doesn't show how the process got the wrong context.

reassigning to selinux-policy.

--- Additional comment from sdake on 2011-09-26 16:57:47 EDT ---

Created attachment 524991 [details]
audit log
Comment 1 Steven Dake 2011-09-26 21:00:58 UTC 
An inability to read a config or operational file (because of selinux for example) should give a more useful error then a parse error.  Instead the filename with an appropriate error should be printed.
Comment 3 Jan Friesse 2012-02-02 09:47:45 UTC 
Created attachment 559020 [details]
Proposed patch

Main problem wasn't hidden in coroparse (parser) but in mainconfig module, which passed incorrect string pointer to function which opens log file. Patch handles this by passing correct string to function, so correct error is now printed.

Reproducer:
set logfile: in corosync.conf to some non-existing directory. Example: logfile: /var/log/cluster2/corosync.log

run corosync.

Old result:
As described in bz

New result:
Feb 02 10:47:24 corosync [MAIN  ] parse error in config: Can't open logfile '/var/log/cluster2/corosync.log' for reason: No such file or directory (2).
Comment 5 Jan Friesse 2012-03-07 07:17:46 UTC 
    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
Cause
Set logging to file if path to file contains nonexisting directory.

Consequence
Error message which doesn't make sense (configuration file error)

Fix
Pass proper string pointer to function for opening log file

Result
Correct error message (can't create log file) is displayed.
Comment 8 errata-xmlrpc 2012-06-20 12:22:37 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHBA-2012-0777.html