Red Hat Bugzilla – Bug 156926
autofs tcp gives errors going to certain netapp boxes
Last modified: 2007-11-30 17:07:07 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7) Gecko/20040616
Description of problem:
We have a number of Network Appliance file servers in our company at various sites around the world. If I enable tcp on autofs and cd via /net to a path on one of these boxes that is not exported in a certain way, I am not able to get to the desired directory. I get a "file not found" error. With udp for autofs I can get to the same directory without a problem. The attached file will show the output from the tcp and udp settings of autofs going to a particular box and the exports file from that netapp box.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Turn on tcp in autofs by editing /etc/auto.net and adding tcp to the options list.
2.Reboot the workstation to make sure the tcp option is functioning in autofs
Actual Results: I get the error "No such file or directory"
Expected Results: I should be able to cd into the directory and do an ls and see the files in the directory.
I will attach a file with output from both tcp and the udp settings going to the same path on the same netapp box. I will also attach the exports file from the netapp box.
I set the Severity level to high because it looks like with the release of Rhel 3 at our site we will have to go with the tcp setting to solve some other issues. This will definately keep our users from getting to at least one of our netapp boxes via /net.
Created attachment 114055 [details]
output from commands and descriptions of the output
This attachment show various commands that were tried and description of the
output that suppliments the original bug.
Please follow the guidelines for submitting autofs bug reports found here:
Also note that the mount commands you do to show that mount works are not the
same commmands that are issued by autofs. Check in the debug logs when you
generate them. Please try the EXACT mount command listed there.
Please also file only one bug per bug report. Open a separate bug to track the
expiration of directories that have references.
Got this update from Dwight:
I went and checked the debug logs and did the mounts exactly as they
were listed in the log file. And sure enough, I got the error "nfs
server reported service unavailable: Address already in use" when I
tried to the manual mount. This error message bothered me because web
research suggests that the error indicates that one of the nfs parties
does not have tcp active. I know it's a part of the kernel in rhel 3
and rpcinfo -p shows both udp and tcp as active on port 2049 on my rhel
3 machine. I had the sysadmin for dfiler4 check to see if tcp was
turned on for this netapp box and it was NOT turned on. I had him turn
on tcp and then I tested the /net mounts back to dfiler4. Everything
So, it appears that this one is on our end with netapp boxes not having
tcp enabled. I had thought that all these boxes were tcp enables by
default but apparently not. This also solved the expiration of
directories as the mount points do not time out from underneath me when
I have cd to one of the directories on dfiler4. I would consider this
I'm closing this as NOTABUG.