Link to home
Start Free TrialLog in
Avatar of sonriks
sonriks

asked on

output of df -k (using automount)

Why are there two entries being returned when I enter the following command?

df -k /nfs/home/oracle/NAS_archoverflow

/nfs/home/oracle/NAS_archoverflow (nasora:/nasarclog     ) : 1013475400 total allocated Kb
                                                  460897032 free allocated Kb
                                                  552578368 used allocated Kb
                                                        55 % allocation used
/nfs/home/oracle/NAS_archoverflow (/etc/nas-map          ) : 1013475400 total allocated Kb
                                                  460897032 free allocated Kb
                                                  552578368 used allocated Kb
                                                        55 % allocation used

There are several map files being used on this server, but this is the only one that is resulting in a duplicate (which references the map file) when I issue the df -k command on this fs.

Avatar of woolmilkporc
woolmilkporc
Flag of Germany image

Hi,

did you over-mount it by hand accidentally (or by /etc/fstab)?


wmp
Avatar of sonriks
sonriks

ASKER

no  .... fstab is clean. Just noticed something though, sometimes the df -k results in both records, and sometimes only one (the one that lists nas-map)
Could there be some sort of procedure (cronjob?) which would mount/demount the fs on a regular basis?
Avatar of sonriks

ASKER

negative .... just checked them all.
Well,

since it looks like Oracle - are you using Oracle's Direct NFS by any chance?

Check for the existence of "$ORACLE_HOME/dbs/oranfstab". Any results?

Depending on the OS flavour you use there might be an /etc/mtab. Please check this as well, it often contains old, invalid mount information.
.. just from curiosity:

do you get the mentioned results with "mount" and "df" alone as well, or just with "df mountpoint"?
Avatar of sonriks

ASKER

Not using Oracle's Direct NFS

There is no  "$ORACLE_HOME/dbs/oranfstab" file

There is no /etc/mtab file

the results I'm getting is from "df -k /nfs/home/oracle/NAS_archoverflow" command

mount command shows only a single record:
/nfs/home/oracle/NAS_archoverflow on /etc/nas-map ignore,direct,dev=4000004 on Mon May  9 15:33:33 2011

Of course, the contents of /etc/mnttab changes .. .sometimes show only one, then other times shows both.

FYI the server was recently rebooted but this condition persists.
That's not consistent -

"the contents of /etc/mnttab changes .. .sometimes show only one, then other times shows both"
versus
"mount command shows only a single record"

??

But I must admiit - no more idea, sorry!

wmp

Avatar of sonriks

ASKER

thanks, wmp .... I appreciate your input.

-s
> There are several map files being used on this server
Could you post the maps?

By the way, do you use autofs?
> (nasora:/nasarclog     )
>  (/etc/nas-map          )
From the df output, it seems that the same mount point show on two maps.

Which Linux system do you have?
Avatar of sonriks

ASKER

Hi wesly,

Here is the auto_master and the map files:

auto_master
/net -hosts -nosuid,soft
/- /etc/crm_map
/- /etc/nas-map
/- /etc/ora-map
/- /etc/vantive_map
/- /etc/sapcd_map

crm_map
/apps/psoft  bizdev2.hns.com:/apps/psoft

nas-map
/nfs/home/oracle/NAS_archoverflow  nasora:/nasarclog

ora-map
/nfs/home/orasoft       nasora:/orasoft

vantive_map
/apps/vantive   bizdev2:/apps/vantive

sapcd_map
/sapcd sapdev01.hns.com:/sapcd

There's no linux server involved here ... things are being mounted from the NAS (nasora)












temporary take out
auto_master
/net -hosts -nosuid,soft   <=== take out this line
...

And reboot this system and check "df -k" command output again.
Avatar of sonriks

ASKER

thanks. I'll give it a try when I'm able to reboot the system. Is there any way to make this take affect without a reboot that might be worth a try?
for Fedora/Redhat/CentOS
as root
service autofs restart

However, this is not clean because some of processes/users might access the NFS mount which the umount will fail.
So reboot give a clean mount and usually flush the nfs cache.
Avatar of sonriks

ASKER

the server in question is an HPUX rx6600. Looking forward to tryout out your suggestion.

Thank you!
HP-UX, check /etc/mnttab for duplicate entries.

This URL suggested that apply Enhanced AutoFS and NFS patches.
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1219285
Avatar of sonriks

ASKER

thanks ... the /etc/mnttab will sometimes show one entry and other times two, just like the df -k

Will be rebooting the server next Tuesday. I'll also check on the patch levels. I'll let you know if the reboot fixes the issue. thanks again for your continued and excellent advice.
Avatar of sonriks

ASKER

The reboot didn't fix the problem. Next step will be to contact HP to see what they say. Thanks again!
Good luck.
Avatar of sonriks

ASKER

Per HP this is normal behavior.
> Per HP this is normal behavior.
If this doen't cause the nfs access failure, then it should be ok.
I saw this in large NFS automount volume environment and I have encountered any NFS issue related to duplicate entries for the same mount point.
ASKER CERTIFIED SOLUTION
Avatar of sonriks
sonriks

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of sonriks

ASKER

This has been decribed as "normal" behavior by HP Support

>> Per HP this behavior is normal. <<

How could it then be that

>> There are several map files being used on this server, but this is the only one that is resulting in a duplicate <<

as written in your Q?

We cannot work seriously on a question when there is wrong or misleading and never corrected information already contained in the original post.