Skip to content | Change text size

IT Services home

Quick links
 

AFS Troubleshooting

If you are having problems during login, it is less likely that it is a problem with AFS. Problems experienced while logged in may be the result of network problems; if not, the following information can be used to tell if there is a problem with AFS.

Checking For AFS Errors

Use fs checkservers to check if AFS file servers are running.

If you are unable to save any files, you may be over quota. Attempts to save files while out of quota will return no space left or similar messages. Use fs lq ~ to check your usage. Delete files (don't forget to empty Trash) to regain quota (quotas are not negotiable).

The most common reason for permission denied errors is expired AFS tokens. Use the tokens command and check the expiry time. If you do not have valid tokens, use kinit to authenticate and then aklog to get new tokens.

AFS Server Software Failures

The AFS system uses a number of server processes on each server. The following sections tell you how to tell if a server is contacteable, and how to tell if server processes are running. These sections assume that you were able to log in.

Check if the problem server is up

From a terminal session, use dmesg to see if there are any recent messages like afs: Lost contact with file server xxx.xxx.xx.xxx in cell cc.monash.edu.au .... If there are no messages like afs: file server xxx.xxx.xx.xxx in cell cc.monash.edu.au is back up ... after it, continue with the next step.

Use ping xxx.xxx.xx.xxx where xxx.xxx.xx.xxx is the same as the numbers reported by dmesg in the previous step. If the server does not respond, use ping sng.its.monash.edu.au to check if you have network failure in your area. If not, then it is probable that the problem AFS file server is down. Check for any downtime announcements before reporting the problem.

If the server is running and the network shows no problems, check if the AFS server processes are running by using bos status xxx.xxx.xx.xxx where xxx.xxx.xx.xxx is the same as the numbers reported by dmesg. Output messages from the bos status command which do not say normally or running should be reported.

Related Information