Errdemon accidently deleted

[KOG]

Folks,

I got an error message on oracle to say that the controlfile was corrupted and the exact nature of this problem is not yet known.

I managed to bring it up online, IBM AIX advised me to remove errlog file and I did it but accidently deleted errdemon too.

Is there anyway I can recover this file, or do I have to reinstall AIX? Any chance of picking it out from CD and transfer it to server?

Regards

Katherine

 

[bi]

Do you have a backup of your /usr/lib directory? That's where errdemon is and it wouldn't change, so even an old backup would do it. After the restore, you would have to restart the daemon.

If you have a mksysb (again, an old one will do), you could restore the file from that. There is a thread in Tek-Tips today that talks about restoring a specific file from a mksysb tape.

The daemon is part of the bos.rte.serv_aid fileset. I would think you would be able to reinstall that fileset (making sure you tell smit to reinstall the fileset), but I'd like to see someone who has actually reinstalled a bos fileset say it's OK.

Good luck.

 

[KOG]

Hi bi,

I will read the info of MKSYSB and hopefully recover the lost file.

Many thanks for your info.

Regards

Katherine

 

[bi]

Before you restore you might want to see if the errdemon is still running in the system's memory (ps -ef | grep errdemon).

Also, you can use smit to restore a single file. Don't know how long it will take, but it can be done. Again, the file you need is /usr/lib/errdemon.

 

[AIXSPadmin]

You can reinstall a bos fileset but you would have to reinstall the base level and do a force install and then apply your maintenance levels. The reason is because bos.rte.serv_aid is already installed, just one file is deleted in the fileset. A reboot would be required after doing the reinstall. I have had problems before with bos.mp and IBM sent a PTF with a new bos.mp and so I went through the force install of base then the new bos.mp. It is safe to do.

 

[KOG]

Hi folks,

Yes errdemon is still running in memory is that good news?

ps -ef | grep errdemon
root 3662 1 0 13 Mar - 0:00 /usr/lib/errdemon
root 50814 15382 1 09:08:22 pts/1 0:00 grep errdemon

To be honest I am a bit reluctant with the idea of doing force installation as I am still a novice when it comes to AIX (I am more comfortable working with oracle).

We are expecting a consultant to come on site for few days next week to assist me in the migration of database from old server to new server, I think I will leave it to him and pray that nothing disaster happens in the meantime.

Regards

Katherine

 

[BFOJ]

Do you only have the one AIX machine? If you have more than one ftp the file from a good working machine to the one missing the file, then most likely having to change the permissions on it. You may have to stop the errdemon process. Then manually restart the errdemon.

What version of AIX are you running, perhaps someone can email you the file.

 

[KOG]

Hi,

I now have two servers (J40 and F50), the J40 is rather old and we have AIX 4.3.2 running on it.

F50 is our new server and I will be migrating the data from J40 to F50 next week with the assistance from consultant. The problem is AIX 4.3.3 is running on F50 which is rather new.

So nothing I can do here really, only restore from last MKSYSB backups?!? Might leave it till Consultant comes next week, just pray and hope all runs well till then.

Regards

Katherine

 

[bi]

Here are my thoughts:

You aren't going to be able to share the files between the two systems because they are at different release levels.

The errdemon file is in the system's memory, so theoretically, it should continue working. But you can't be sure it will stay in memory, although since it is something that is constantly monitoring your system, AIX probably does keep it in memory. If that is the case, it will stay there and continue working until the system boots.

When the system boots, if you don't have the file replaced, you may receive an error because the /sbin/rc.boot file tries to start the errdemon, among many other things. (rc.boot is the third and final phase of the boot process and is invoked from the second line of your /etc/inittab.)

If the maintenance level of the system when you made the mksysb tape was the same as what your system is at now, I would think there would be no problem in restoring the errdemon file. Are you sure that is the only one that was deleted?

I would stop or kill the errdemon process, do the restore and then restart the process.

I have just performed a quick and dirty test on a sandbox here. I copied the errdmon file to another filesystem, then deleted it from /usr/lib. Then I forced an error by filling up a test filesystem. The error was logged. I copied the errdemon file to /usr/lib from the filesystem I had copied it to and restarted the demon (by typing /usr/lib/errdemon). Then I forced another error by filling up another test filesystem. It worked.

I think the important thing to do is to make sure the ML of the system on the mksysb tape is the same as what you have on the system now.

And, by the way, a force install isn't as bad as it sounds. You just have to tell AIX to go ahead and reinstall that particularl fileset even if it is the same or older than what is already on the system. Then, after the installation, you would have to reapply the ML to get that group of filesets up to the same level as the rest of the box.