View Issue Details

IDProjectCategoryView StatusLast Update
0006734CentOS-6autofspublic2013-11-12 20:29
Reportersabujp 
PrioritynormalSeverityminorReproducibilityalways
Status newResolutionopen 
Platformx86_64OScentosOS Version6.4
Product Version6.4 
Target VersionFixed in Version 
Summary0006734: autofs mounts don't resolve ip's in the output of df for round robin dns entries
DescriptionThis is similar to bug 5109. Autofs mounts aren't showing the short version of the hostname that is mounted but instead the IP address. The system can resolve the reverse. Here's what df -h shows :

10.96.193.21:/dors/csb/dist/
                      5.0T 3.4T 1.7T 67% /sb/dist

It should show :

mako:/dors/csb/dist
                      5.0T 3.4T 1.7T 67% /sb/dist

[root@scylla ~]# host mako
mako.mc.vanderbilt.edu has address 10.96.193.22
mako.mc.vanderbilt.edu has address 10.96.193.21
mako.mc.vanderbilt.edu has address 10.96.193.20
[root@scylla ~]# host 10.96.193.22
22.193.96.10.in-addr.arpa domain name pointer mako.mc.vanderbilt.edu.
[root@scylla ~]# host 10.96.193.21
21.193.96.10.in-addr.arpa domain name pointer mako.mc.vanderbilt.edu.
[root@scylla ~]# host 10.96.193.20
20.193.96.10.in-addr.arpa domain name pointer mako.mc.vanderbilt.edu.
[root@scylla ~]#
Steps To Reproducesetup autofs to mount a host that has round robin dns entries
TagsNo tags attached.

Activities

sabujp

sabujp

2013-11-12 15:46

reporter   ~0018336

autofs-5.0.5-75.el6_4.x86_64
toracat

toracat

2013-11-12 17:25

manager   ~0018337

From what I gathered, that (showing the IP instead of the hostname) is the intended behavior when the hostname has multiple DNS addresses.
sabujp

sabujp

2013-11-12 17:44

reporter   ~0018339

No, because it doesn't work like that on other systems with older versions of autofs (e.g. centos 5.6, autofs-5.0.1-0.rc2.143.el5_6.2) . If I want the IP I can run mount . Actually after 5.6, i.e in 5.7 was when the problem started happening (or behavior changed) according to bug 5109.
sabujp

sabujp

2013-11-12 17:46

reporter   ~0018340

in any case, I'll check upstream ..
toracat

toracat

2013-11-12 18:15

manager   ~0018341

Again from what I gathered, looks like the current behavior is new in EL6. Older autofs probably does not have this "feature".
sabujp

sabujp

2013-11-12 19:10

reporter   ~0018343

Just tried autofs 5.0.8, does indeed look like they changed this behavior. I'll go find out why.
toracat

toracat

2013-11-12 19:19

manager   ~0018344

CentOS must match upstream, feature for feature, bug for bug. So, I suggest you file a report at http://bugzilla.redhat.com and see what they say about why the current behavior was made default.
sabujp

sabujp

2013-11-12 19:29

reporter   ~0018345

By upstream i meant autofs devs, but not sure if the autofs devs are also rht devs. I'll check.
toracat

toracat

2013-11-12 20:29

manager   ~0018346

OK, keep us posted with your findings.

In the meantime I spotted the following entry in the changelog (last line):

* Fri Jan 28 2011 Ian Kent <ikent@redhat.com> - 5.0.5-27
- bz616426 - autofs initscript should implement force-reload and try-restart
  - add lsb force-reload and try-restart
- bz629359 - [RFE] Automounter debug mode should include IP addresses
  - include ip address in debug logging.
  - mount using address for DNS round robin host names.

Issue History

Date Modified Username Field Change
2013-11-12 15:42 sabujp New Issue
2013-11-12 15:46 sabujp Note Added: 0018336
2013-11-12 17:25 toracat Note Added: 0018337
2013-11-12 17:44 sabujp Note Added: 0018339
2013-11-12 17:46 sabujp Note Added: 0018340
2013-11-12 18:15 toracat Note Added: 0018341
2013-11-12 19:10 sabujp Note Added: 0018343
2013-11-12 19:19 toracat Note Added: 0018344
2013-11-12 19:29 sabujp Note Added: 0018345
2013-11-12 20:29 toracat Note Added: 0018346