Unstable Shared Folder

M

MikkoL

Bit poster
Hello,
I've used Parallels 6 + Ubuntu Maverick for quite long successfully, but since upgrade to Natty, alongside with the update to parallels tools to support it, I've gotten problems with my use of Shared Folders.

I got in to habit of developing on my mac side editor, and then doing test compiles on the linux side, through the shared folder, and it worked perfectly. But after upgrading, I've experienced a regression, with messages like these:
make: stat: probe/main.o: Stale NFS file handle
cc -std=c99 -Iinclude -Wall -Wextra -pedantic -ggdb -Werror -c -o main.o main.c
cc: Internal error: Killed (program as)
Click to expand...
Doing a `find .` in the same directory sometimes results in:
find: File system loop detected; `./include/ngi' is part of the same file system loop as `./include'.
Click to expand...

So, the workflow I got accustomed to isn't usable at all.

I have tried removing and reinstalling Parallels Tools.
 
similar "Stale NFS file handle" problem

I also have a similar problem with Parallels 6 and my Ubuntu virtual machine. In my case I am editing files located in my mac shared folder, using Eclipse from Ubuntu (I am using one plugin that only works fine in Ubuntu). In certain cases I am getting the Stale NFS file handle error message when I try to create/delete files from Ubuntu.

thanks for any help,

Sergio
 
Seeing the same behavior in Natty as both above have said, even after the most recent patch that I pulled down a couple hours ago.
 
What's the status on this Parallels? I use Parallels Desktop for work daily so this is a big issue for me.
 
Please fix

I too have been experiencing this issue consistently since updating to Ubuntu 11.04

When I run an ls -l on a directory accessed across the share, I get:

ls: cannot access image.jpg: Stale NFS file handle
total 0
drwxrwxrwx 1 root root 306 2011-10-17 21:18 ./
drwxrwxrwx 1 root root 204 2011-07-07 22:10 ../
?????????? ? ? ? ? ? image.jpg
 
same here.....

Moved from VMWare to Parallels but not too happy about it. I have the exact same issues with the shared directory and NFS stale handles.
 
NFS Handle Stale

Just updated to 7.0.15050 and same problem with Linux Ubuntu/Fedora (NFS handle stale).
 
Seven months, still a bug

Same issue, "File system loop detected" in 7.0.15054. Tried to open a ticket to ask for help, but they just closed it and said a ticket is not needed. It was unclear if they intended to work on the issue, as they never acknowledged it as a problem in any way.

In the mean time, shared folders are remain essentially unusable.

I've looked at alternatives ranging from CIFS/Samba to NFS, but nothing seems to work that well in a mobile environment without a good deal of time on the command line each time I need to start working.

It's been seven months, I think we can assume they just gave up, so anyone have a good alternative to recommend? I have had good lunch with SSHFS on the Linux side.

http://www.wiredrevolution.com/system-administration/share-a-remote-filesystem-over-ssh
 
Last edited:
Workaround corrupted filesystem prl_fs "file system loop detected"

This "File system loop detected" problem is something that affects any user (so don't be looking at your personal configuration; it won't help). I managed to get a ticket for this with Parallel's support, but once it was recognized as a problem that the front-line support could not fix, they marked it as "resolved" and said that they had passed it on to the their development team... (so, not really "resolved").

But, some of us cannot wait for the Parallel's development team to fix bugs, so we have to look for alternative ways to share files between the host MacOS machine and a guest VM. The easiest way is to use NFS. It might not provide rapid access to the files, but it works (no more file system loops).

On the host MacOS machine, create a file /etc/exports containing the following:

/Users -alldirs -mapall=501:20 -network 10.37.129.0 -mask 255.255.255.0
/Users -alldirs -mapall=501:20 -network 10.37.55.0 -mask 255.255.255.0

The IP addresses after the network option are those of the subnets defined for shared and host-only connections. You can find your values in Parallels Desktop, Preferences..., Advanced, Network: Change settings..., see Subnet: value.
Access to shared files is thus limited to the subnets of your virtual machines, which should address any security issues for the majority of us.

Once the /etc/exports file has been created, just reboot the machine. It's the easiest way to get the nfs server up and running.

For an Ubuntu guest VM, edit the /etc/fstab file, adding the following:

10.37.55.2:Users/macosUserName /mnt/host nfs auto,sync,noatime,nosuid,exec,nodev 0 0

where the 'macosUserName' and the IP address of the host machine in your local subnet (here 10.37.55.2) is to be changed to your needs (replace the final .0 of the subnet IP with .2)

sudo mkdir /mnt/host
sudo mount /mnt/host

and you're done.

find /mnt/host -name "toto" -print
Ouuh, that's nice, it works.

Note NFS is supported only by some editions of Windows (win7 enterprise edition and above).
 
You must log in or register to reply here.
Back
Top