How to umount a busy NFS mount in solaris 2.6?

How can I umount a busy NFS mount in solaris 2.6 without rebooting the machine? I've tried to check what processes that might hold the mount with the command 'fuser -c /home/user' but it reports nothing.
I wish I had solaris 8 so I could use 'umount -f'...
Can anyone help me?

If your nfs mount is under autofs' control, you would run into issue. Give this a try.
# /etc/init.d/autofs stop
# fuser -ck /home/user
# umount /home/user

Similar Messages

How to change default Server NFS version in Solaris 10 ?

How to change default Server NFS version in Solaris 10 ?
I need to change default version of NFS server version, currently i have NFS_SERVER_VERSMAX=2, for changing NFS version i changed this variable in /etc/default/nfs, but still its using version 2,
Please let me know where else i need to change./
Thanks a lot in advance!!!

Did you change the /etc/default/nfs on the server or on the client?
What is NFS_SERVER_VERSMIN set to? The NFS server is allowed to use any version in the span of NFS_SERVER_VERSMIN and NFS_SERVER_VERSMAX.
NFS_SERVER_VERSMAX=4
The default of NFS_SERVER_VERSMIN is 2.
And, as mentioned, you will have to restart your NFS server for changes to take effect.
.7/M.

Permanent NFS Mount

Hi
I'm trying to figure out how to create a permanent NFS mount on my 10.5.6 Server hosts. Using Directory Utility seems to only create autofs mounts - which I've had trouble with in the past on other platforms, so I'm not very trusting of it.
/etc/fstab.hd is apparently ignored, so I'm not sure how else to get a permanent mount. Is it even possible on Leopard Server?
Thanks.

I also found that nested / heirarchical mounts don't work with the Apple version of Sun's automounter. I thought for sure if I made a ``multiple mounts'' entry they would work since AIUI in this case the automounter gets to see the whole proposed subtree at once from a single Directory Services lookup, but no, it doesn't work either. There are ``multiple mounts'' examples on both Apple's man page and Sun's, but on Sun's page the example is nested and on Apple's it isn't. I guess that's a kind of transparency, but a rather CYA-ish kind that leaves us out here wagging our jaws quite a bit when we expect it to behave like other automounters.
However! nested mounts with the 'net' option DO work. ?!
and in this case unlike the traditional Sun ``multiple mounts'' case, the automounter must build the tree with multiple Directory Services lookups not just one. How can it even do that? Is it Searching the directory instead of doing a simple Lookup?
I can load all the mounts into Open Directory as separate, nodes, or whatever you call them, like this:
cat > nested-example
0x0A 0x5C 0x3A 0x2C dsRecTypeStandard:Mounts 3 dsAttrTypeStandard:RecordName dsAttrTypeStandard:VFSType dsAttrTypeStandard:VFSOpts
terabithia\:/arrchive/incoming:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/Radio:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/backup:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/ebooks:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/fonts:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/movies:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/music/Antoine:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/music/Lauren:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/music/Roger:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive/music/jen:nfs:nosuid,nodev,hard,intr,net
terabithia\:/arrchive:nfs:nosuid,nodev,hard,intr,net
^D
dsimport -g nested-example /Local/Default I -u someadminuser
and they will show up under /Network/Servers/terabithia/arrchive. I can no longer choose the mountpoint myself, which is a disadvantage for more than vanity---with the Solaris automounter, it's possible to build a single nested tree on the client out of filesystems pulled in from a bunch of different NFS servers, while the Mac's 'net' naming convention straightjackets me into only rebuilding trees that exist within one NFS server.
Also, this works on 10.4, too, though in that case of course you use netinfo or niload fstab instead of dsimport.
Now can someone explain why 'net' suddenly works so much better? And is there a hidden downside to using it?

Autofs timeout while accessing to remote NFS mount

Following Apple recommendations, I switched to "Directory Utility" to configure NFS mounts. As far as I understand, if you do so, the mounts are handled by automount which is itself called by autofs. The good thing of this is that autofs is unmounting unused mounts (after a timeout of 3600 seconds as defined in /etc/autofs.conf). Any time you need the remote drive (Finder call, ls in Terminal..., opening file), autofs is remounting ressource. This is a nice behaviour... in theory.
I'm using some codes (written in IDL) that are reading and writing on that remote NFS server once every 5 minutes. Theoretically, autofs should be detecting these accesses and should keep the drive mounted. This is unfortunately not the case. The drive is unmounted 3600 seconds after I last accessed the mount through the Finder or with any other Application.
There is apparently no way to remove this "automated unmounting" feature. I tried to set the timeout delay to a very large number (1 day) but it still disconnects me after this delay, if I don't do anything else than running my IDL code. If I mount the NFS share with the "mount_nfs" command, it works perfectly, as it is not handled by autofs.
I wonder then if there is any recommandation on Apple's side in such a case, other than going back to traditional mount_nfs.

As you have discovered, automount/autofs is also an "auto-unmounter" and there is no way to remove that feature. Contrary to what one might think, the auto unmounting does NOT happen after a period of "inactivity" of the mount. This is because autofs has no way of knowing when an automounted file system was last accessed. So, instead it periodically attempts to unmount it - if it is busy it won't get unmounted - if it isn't busy it will get unmounted.
You can't disable this - but you can make the periodic unmounting so infrequent as to effectively disable the feature. Try setting the AUTOMOUNT_TIMEOUT interval to something really large - like 315360000 (which would be 10 years).
However, in theory, this auto-unmounting should not be a problem because if it does get unmounted then the next access to that file system should cause it to get mounted again. And all this should happen without the code that is accessing the automount knowing that it isn't always mounted. It should always be there when it is accessed. So, the usual response to someone asking how to disable the auto-unmounting is to ask why they think it is a problem.
(Oh, and you don't have to use "mount_nfs" - just "mount" should work to manually mount an NFS file system (that saves a little typing).)
HTH
--macko

Systemd and nfs .mount

Hi,
I'm trying to setup a systemd unit for my NFS mounts. I don't want to use /etc/fstab because I want an openvpn.service dependency. However, for now, I'm omitting that dependency due to debugging. Here is the current unit file I have:
# cat host\@.mount
[Unit]
Description=%i mount
DefaultDependencies=no
Requires=local-fs.target network.target rpc-statd.service
Conflicts=umount.target
[Mount]
What=host:/%i
Where=/%i
Type=nfs
Options=user,async,atime,exec,rw,wsize=32768,rsize=32768
DirectoryMode=0755
TimeoutSec=20
[Install]
WantedBy=multi-user.target
When I try to enable this I get the famously vague error:
# systemctl enable ./host\@mnt.mount
Failed to issue method call: Invalid argument
Any ideas on how to fix this? Thanks!

The solution is NOT to create this file at all. Apparently, exports from the server do not require them. If I remove it and reboot the server, I am able to connect from my workstation with no issues. For reference:
$ ls -l /etc/systemd/system/multi-user.target.wants/
total 0
lrwxrwxrwx 1 root root 40 May 10 10:58 cpupower.service -> /usr/lib/systemd/system/cpupower.service
lrwxrwxrwx 1 root root 38 May 10 10:58 cronie.service -> /usr/lib/systemd/system/cronie.service
lrwxrwxrwx 1 root root 40 May 10 12:10 exportfs.service -> /usr/lib/systemd/system/exportfs.service
lrwxrwxrwx 1 root root 42 May 10 10:59 lm_sensors.service -> /usr/lib/systemd/system/lm_sensors.service
lrwxrwxrwx 1 root root 35 Apr 30 15:15 network.service -> /etc/systemd/system/network.service
lrwxrwxrwx 1 root root 36 May 10 10:59 ntpd.service -> /usr/lib/systemd/system/ntpd.service
lrwxrwxrwx 1 root root 36 May 10 11:33 rc-local.service -> /etc/systemd/system/rc-local.service
lrwxrwxrwx 1 root root 40 May 2 22:37 remote-fs.target -> /usr/lib/systemd/system/remote-fs.target
lrwxrwxrwx 1 root root 39 May 10 10:58 rpcbind.service -> /usr/lib/systemd/system/rpcbind.service
lrwxrwxrwx 1 root root 42 May 10 12:10 rpc-mountd.service -> /usr/lib/systemd/system/rpc-mountd.service
lrwxrwxrwx 1 root root 41 May 10 12:10 rpc-statd.service -> /usr/lib/systemd/system/rpc-statd.service
lrwxrwxrwx 1 root root 43 May 10 10:58 sshdgenkeys.service -> /usr/lib/systemd/system/sshdgenkeys.service
lrwxrwxrwx 1 root root 36 May 10 10:58 sshd.service -> /usr/lib/systemd/system/sshd.service
lrwxrwxrwx 1 root root 41 May 10 11:06 syslog-ng.service -> /usr/lib/systemd/system/syslog-ng.service
lrwxrwxrwx 1 root root 35 May 10 10:57 ufw.service -> /usr/lib/systemd/system/ufw.service

Vi error on nfs mount; E212: Can't open file for writing

Hi all,
I've setup a umask of 0 for testing on both NFS client (Centos 5.2) and NFS server (OSX 10.5.5 server).
I can create files as one user and edit/save out as another user w/o issue when directly logged into the server via ARD.
However, when I attempt the same from an NFS mount on a client machine, even as root I get the following error using vi;
"file" E212: Can't open file for writing
Looking at the system.log file on the server, I see;
kernel[0]: add_fsevent: no name hard-link! dropping the event. (event 2 vp == 0xa5db510 (-UNKNOWN-FILE)).
This baffles me. My umask is 0 meaning files I create and attempt to edit as other users are 777, but I cannot save out edits unless I do a wq! in vi. At that point, the owner of the file changes to whomever did the vi.
This isn't just a vi issue as it happens using any editor, but I like to use vi.
Any help is greatly appreciated. Hey, beer is on me!

Hi all,
Thanks for the replies
I've narrowed it down to a Centos client issue.
Everything works fine using other Linux based OS's as clients.
Since we have such a huge investment in Centos, I must figure out a workaround. Apple support wasn't much help as usual however they were very nice.
There usual response is "its unsupported".
If Apple really wants to play in the enterprise of business space, they really need to change there philosophy. I mean telling me that I shouldn't mount home directories via NFS is completely rediculus.
What am I supposed to use then, Samba of AFP? No, I don't think so. No offense to Microsoft but why would I use a Windows based file sharing protocol to mount network shares in a Nix env???

Cannot write to NFS mount with finder

hi folks
i have moved from a G4 cube running 10.4.9 to a new iMac running os x 10.5.7 (including upgrading from the old machine)
the NFS mounts i used to have with my G4 cube are not working properly; new mounts i've created aren't working properly either.
i can read/open file OK, but when i try to drag & drop files onto the NFS mount using the finder i get errors:
"You may need to enter the name and password for an administrator on this computer to change the item named test.jpg" [ stop ] [ continue ]
clicking "continue" i get:
"The item test.jpg contains one or more items you do not have permission to read. Do you want to copy the items you are allowed to read? [ stop ] [ continue ]
Choosing continue again results in the file appearing in the NFS directory, but with 0 size, and a time stamp of 1970.
if i try to copy the same file using the Terminal, it works fine - so it is not a simple NFS permissions problem - it is something particular to the Finder.
i am able to create a folder inside the NFS director by using the Finder.
i thought at first it might be related to the .DS_Store and similar files being written, so i tried turning off that behaviour:
defaults write com.apple.desktopservices DSDontWriteNetworkStores true
but that hasn't fixed the problem
there are no obvious messages in any of the logs
any suggestions or pointers on how to fix this?

thanks for the reply
these articles appear to relate to sharing a mac filesystem via NFS: exporting the data.
i am referring to mounting a NFS filesystem from another server onto the mac (leopard) client
the mounting works fine: it's just the finder which isn't behaving. the finder worked in tiger; isn't in leopard.

Missing menu "File|NFS Mounts" in Disk utility

I had many NFS mounts using Disk utility and I used "File|NFS Mounts" but now that option is missing and I can't see my mounts neither those mounts are working, so I have two questions
1. Where I can see my old mounts listed?
2. How can I make nfs mount work?

The problem is Boot Camp: It uses a hybrid GPT/MBR partitioning scheme - which ends up hiding the Recovery HD partition - which is an EFI physical partition (neither GPT nor MBR).
I would expect a new version of Boot Camp to release - like real soon - because of the Recovery HD partition invisibility issue.
In this article - it is suggested that rEFIt should be used to partition a hard drive that is going to support multiple boots including Mac OS X, LINUX, and Windows.
(http://wiki.onmac.net/index.php/Triple_Boot_via_BootCamp)
The key in the article to using Boot Camp with rEFIt is this:
"Run the Boot Camp Assistant and create the Windows XP driver cd. Then exit Boot Camp. DO NOT PARTITION USING BOOT CAMP: you are only using Boot Camp for the drivers, not the partitioning."
All partitioning is done in terminal mode using the "diskutil" command.
rEFIt is used to update both the GPT and MBR records so that all partitions will be visible using its "gptsync" command.
Then - you replace the standard Mac boot menu with the rEFIt boot menu. THAT will show the Mac OS X partition, Recovery HD (an EFI partition), and the Windows partition.
My caveat is that rEFIt - which is open sourced and available here: http://refit.sourceforge.net
has not been recently updated and tested with respect to Mac OS X Lion.
Hope this helps!

IPhoto 09 & storing "Originals" directory in NFS-mounted partitition

In iPhoto 08, I had the Originals directory NFS mounted off a Linux server. On the Mac, I mounted it at /Volumes/pics and I had a symbolic link pointing from iPhoto Library/Originals to /Volumes/pics. Everything was hunky-dory.
I upgraded to iPhoto 09 and now it will not import new pictures. I get the following error message:
iPhoto cannot import your photos to this library because iPhoto cannot access the library.
Anyone have any ideas on how to store the original pictures on an NFS-mounted partition

Ok. I figured this out myself.
I mounted the NFS volume directly on ~/Pictures/iPhoto Library/Originals. This seems to fix the problem.
So, it's not the NFS mounting that's the issue. It's the symbolic link to /Volumes/pics.
Weird.

Setting umask on NFS mounts under OS X 10.6

I don't see any up to date information on setting umasks with 10.6. We use NFS SAN storage for several folders, mounted using the new NFS Mounts feature of Disk Utility and the SAN and Mac are playing nicely for UID and GID mapping even using our Active Directory LDAP groups and accounts.
However, it creates everything as 755 (umask 0022, the system default) and I'd like to change that umask to 002 to create 775's instead. Nothing for 10.4 or 10.5 seems to apply here (NSUMask, etc.) and it doesn't appear to be a NFS Mount option, either.
NFSv4 and ACLs are not really a good option for me since it's a multiprotocol SAN environment we want to keep NFSv3 permissions bits for UNIX/Linux/MAC and ACLs for Windows. Any ideas how to set the default umask, or at least a way to set the umask for each mount?

xzi wrote:
Yes, I tried that, it does not appear to work on 10.6 that's my point.
I'm sorry, but I just can't figure out "etc". All I can do is keep asking if you have tried this or that. Have tried both variations of that config file and it didn't work? Have you tried instructions in this document, which seems updated for Snow Leopard?
Have you considered that you may have already tried the correct solution, but simply didn't do it correctly? Have you tried with logging-out and back in? With rebooting?

Global NFS Mount & Cluster

Dear All
My Development server is in LAN environment and other system QAS and PRD is in the SZ2. For transport management configuration, we need to do the global NFS mounting. But as per my company policy, there is security issue.
The Second issue is that if we mount /usr/sap/trans as global and also part of the NFS then it cluster startup will fail. Please suggest the above directory should be part of cluster or not.
Regards
Vimal Pathak

Tiffany wrote:
          > We need to store information (objects) that are global to a cluster.
          The only way you can do this is to store the information in a
          database.
          > It's my understanding that anything stored in the servlet context is
          > visible to all servers,
          No. This is not true.
          > but it resides on a network drive. Wouldn't
          > each read of this servlet context info involve a directory read hit
          > with all its implied performance degradation?
          How about WebLogic Workspaces? Is this information replicated across
          clusters? Does it live on a network drive as a file as well?
          > Hoping someone can help us out here.
          Workspaces are not replicated.
          >
          > Thanks for any help,
          >
          > Tiffany
          Cheers
          - Prasad

Backing up to an NFS mount

Has anyone figured out a way to backup to an NFS mount? Ideally, I'd like to create something like a 500GB sparse bundle disk image on /Network/Servers/foo/Backups/my_machine.sparsebundle, and just have time machine write it's backups to this image, similar to how it works for AFP mounts. I can't seem to figure out how to get a disk image to show up as a valid TM target disk though.

There seems to be an unsupported solution (search for "time machine nfs" with you favourite search machine). Nevertheless it does not (yet?) work.

Removing desktop icon of NFS mounted drive

Hi,
I have an NFS mounted drive (on a linux server) on my mac mini, running OS 10.4. I don't want the icon to appear on the desktop. Is there any way to remove this?
The reason is as follows:
- the NFS mounted directory is of the form servername:/home which contains home directories for all the students. This is mounted to /Volumes/servername. The students home directories are set to /Volumes/servername/userid. Thus, in the finder, their home directory shows up correctly, and they can browse their network files.
- However, /Volumes/servername is shown as a mounted disk icon on the desktop, and if students open this, they get a list of everyone's home directories, which they have to scroll through to reach their own home directory. This gives them immediate access to other people's home directories. Granted, these should all be locked and inaccessable to them, but if anyone has messed up their permissions, everyone will instantly be able to see that there is no red X by the home directory, and they can browse in.
- I realise people can browse in anyway, through the Finder, but its just nice not to make it quite so obvious to them.
So... any ideas on how to remove this desktop icon? Note that it needs to be removed for all users (a hundred or so students, logging in through LDAP).
Thanks!
Richard Grant
Mac Mini Mac OS X (10.4.3)

I have figured this out now...
If you mount the directories at the point /home (or at least ask it to - in fact it mounts it somewhere in /private) then the desktop icon is not displayed by default.
Richard

Solaris Zones and NFS mounts

Hi all,
Got a customer who wants to seperate his web environments on the same node. The release of apache, Java and PHP are different so kind of makes sense. Seems a perfect opportunity to implement zoning. It seems quite straight forward to setup (I'm sure I'll find out its not). The only concern I have is that all Zones will need access to a single NFS mount from a NAS storage array that we have. Is this going to be a problem to configure and how would I get them to mount automatically on boot.
Cheers

Not necessarily, you can create (from Global zone) a /zone/zonename/etc/dfs/dfstab (NOT a /zone/[i[zonename[/i]/root/etc/dfs/dfstab notice you don't use the root dir) and from global do a shareall and the zone will start serving. Check your multi-level ports and make sure they are correct. You will run into some problems if you are running Trusted Extensions or the NFS share is ZFS but they can be overcome rather easily.
EDIT: I believe you have to be running TX for this to work. I'll double check.
Message was edited by:
AdamRichards

Stop scanning/mounting NFS mounts

Hi,
I have a Linux desktop. I used to transfer files b/w my MacBook and my desktop using NFS. I don't require to transfer files anymore and hence I stopped using the NFS service a long time ago. But my Mac still searches/tries to connect to the server, and my desktop HDD still shows up (looks like an alias) in the various disk monitoring applications. Is there a way to stop this? The Console message is as follows:
com.apple.automountd: mount_nfs: can't mount /harshad from 192.168.1.4 onto /Volumes/minty: Host is down
And another completely unrelated question: How do you stay signed in in the support communities?
Thanks.

You need to do a "svcadm enable" for the following services, or NFS mounts at boot won't work...
svc:/network/nfs/status
svc:/network/nfs/nlockmgr
svc:/network/nfs/client

How to umount a busy NFS mount in solaris 2.6?

Similar Messages

Maybe you are looking for