AIX Login issues...
 

Greetings everyone!

I have a box at work that is running AIX version 5 and over the last month or so we have been having recurring issues with bringing this box online after a restart for maintenance. Usually it would come up to a login screen and we would login with our common user name, wait for a bit of type to scroll, and then hit the enter button and we were on our way. Lately though it has gotten to where when the login screen comes up we get the message that we must "exec" login from the lowest login shell when we try to login with any user name even su.

I suspect there is a corrupt file somewhere but not being familiar enough with AIX I am uncertain how to proceed. When the box comes up we get these messages now:

export fs scitex/version/pfcProfilesActions: file or dir does not exist

and then a bit later this comes up-

cannot create datavolumes/Brisque07.1.0
cannot create datavolumes/Brisque07.1.0 not enough space in the system

and a little later-

failed write of utmp entry rctcpip
failed write of utmp entry rcnfs

In the past successive restarts of the box were successful in eventually bringing it back online but this no longer appears to be working not that that was ever the desired approach to begin with, it was just one a few of my co-workers chose to adopt. I'd prefer to address the problem at its root (no pun intended) and tackle the matter head on, get it sorted, and be prepared should the same thing happen again in the future.

I don't have a lot of UNIX experience. I'm the first to admit that but I do know my way around computers and I'm a quick study if I do say so myself.

I'd appreciate any assistance.

Just wanted to add that after posting the above I finally got it to accept my root login on about the fifth or sixth restart. I'm running scismit now and one of the first messages up is-

mv: //rc.net.out: No space left on device

I'm not exactly sure what that means but it doesn't sound good because this message and the ones in my previous post have only started showing up on our monitor in the last week or so which suggests to me the corruption is getting worse.

Typically after scismit runs we can log in with our common user name but I would like to eliminate all of this pre-login annoyance if possible.


UPDATE:

After running scismit I was able to login with our common user name. I would still appreciate any help with this problem though.

Hello,

Looks like you're getting lots of 'no space left on device' errors, so that should be your first step in cleaning up your system. Execute:
to find out which partitions/disks are out of space and clean them up.

Kind regards,

Eric

Ok I have done that. Is there a way to save the terminal report to a text file that I can paste here?

Hello,

Yes, by using redirection like this:
and change filename.txt to whatever you want. How much experience do you have with Linux/Unix? As what user are you logged in, root, regular user?

Kind regards,

Eric

Regrettably only a little experience working from the command line. My work on our UNIX boxes apart from running the maintenance script has largely been in the GUI of our RIP program. However I had managed to figure out that what you suggested was the path I needed to tread but I wasn't using proper UNIX syntax! :-) Let me try your suggestion there.

Right now I am logged in as a regular user but I can log in at the root level if need be.

My shift is nearly over. I'll post more information when I come back on tonight. Hopefully I'll "speak" to you later.

Hello,

OK, talk to you later, let's see if we can get your problem fixed later on.

Kind regards,

Eric

I'm working within the terminal the Brisque client gives me at the moment. As far as I know the only way to get to the actual terminal at UNIX level would be to restart the box and I don't care to go through all of the headache of restarting and logging in until it finally accepts my login.

When I use the command above in the Brisque terminal it just says command not found. There may be a way to print from the Brisque terminal but I can't find it.

The df-k command does say that the dev/hd4 is 100% used though. I don't really see anything else unusual but I do wish I were able to post the report.

Hello and good morning,

What's the 'Brisque' client you're referring to? Is it like an SSH client software you use to connect to AIX? If df -k shows you that hd4 is at 100% then I'd start troubleshooting there. Check the mountpoint for that disk by typing:
You should get some lines like this:
Mind you these are from a Linux box, not from AIX but you should be able to find out where it's mounted. It's holiday here in Spain and I don't have access to our AIX servers at the moment, so I'm unable to check the exact output.

Kind regards,

Eric

The Brisque client I'm referring to is the RIP software we use between our Mac workstations and the UNIX boxes that basically translate our files from "Macspeak" to "UNIXspeak" so I suppose it would be functioning as the SSH client.

When I type mount it tells me that /dev/hd4 is mounted over /. Does that help any?

I hope you're enjoying the holiday!

Hi,

What's the size of the disk/partition? Do you have /var in there too? Or is /var mounted on another disk/mountpoint? First thing you could do if /var is included in the / disk, is check the log files in /var/log and delete any old logs and see if that frees up some space.

I am enjoying the holiday, thanks.

Kind regards,

Eric

I have a /dev/hd9var mounted over /var. Is that it? The df-k command tells me that dev/hd4 reads 131072 under the 1024 blocks column.

Does that tell you anything? How would I access those logfiles to clean them out?

Hi,

OK, so the /var is assigned to another disk/partition. Then you'll have to look in your / for 'big' files that are obsolete and remove them. Be careful though with what you delete because you can easily break your system. Before all: do you have a good and up to date backup of your system (mksysb)?

Kind regards,

Eric

Yes we do have a backup.