I have Two 8203-E4A'a (P6 520), identical hardware purchased at the same time. Once server has performed flawlessly. The other server, build from a mksysb from the machine it replaces, has had to be restored 5-6 times, from an unrecoverable 0517 /usr error, and rootvg is on mirrored drives.
The IBM tech has run all the diags, booted off the diag CD's, and no errors are found. The box can run fine for months, and then we will get this error. The most recent error occured morning after we updated to AIX 5.3 TL11 SP4. We updated 10 other servers to the same oslevel, and no issues.
Should I restore the mksysb again(the /usr recovery has never worked); more hardware tests; or start with a fresh OS install
dukessd 2000004Q8P345 Posts
Re: Repeated 0517 LED error2010-09-28T20:37:03ZThis is the accepted answer. This is the accepted answer.You have the choice.
You can either give up on the bad software now and reload from scratch (Hardware support have not found a problem and these systems are pretty good at reporting hardware problems) or you can raise an AIX software call with IBM. They should talk you through a debug boot to find out where the boot process is failing. Then they should be able to talk you through a maintenance boot from NIM / CD / mksysb so you can run fsck and also grab a copy of any failing filesystems BEFORE they are repaired so that the corruption can be analysed.
What attempts have you made in the past to recover from LED 517 other then a mksysb restore?
Does it fail on every reboot?
Have you tried:
The problem with the above doc is that although it solves the boot problem it also destroys the evidence so you may face the problem again if there is another gremlin in the machine.
I'd call IBM AIX support and try and get to the root of the problem, if it is repeatable you can send them a mksysb and they can recreate the problem in-house and then let you know the answer.