rpm database hosed

[DigiCrime]

With it being damaged cpanel updates cant complete. short of rpm --rebuilddb and removing the lock files and doing a rebuilddb again its not going to work.. any one have any other ideas on how to repair it (aside from a system reload) ? I contacted cpanel support about this to get their advice and input they said the drive was going bad which is what caused it to stop working. Logically that doesn't seem right as you would see other failures...i havent found any evidence that the drive is failing that would cause this anyone else had this happen?

System is already hosed what if i transfer an RPM database from a duplicate server, their the same in software (im thinkin no here) but im all for trying the ludicris to. sure there would be difference in signitures but cant hurt it anymore then it already is

 

[chirpy]

That's your only realistic option if you want to avoid an OS restore - copy the database from another server with the same OS and revision and update level. As you say, there will be differences, but they're usually manageable and you're better off that you were. Copy the whole of /var/lib/rpm/

 

[DigiCrime]

Didnt help, oh well os restore not a big deal just not following their logic on this. They see some driveseek errors in message log and the rpm damage is caused by this so they say... however I have systems that have shown these errors for a couple years or longer and still present to this day that dont exhibit this kind of error/issue if the drive was going bad hdparm would show some slow readings probably or other services would be failing or far worse.

Whats your opinion on this?

 

[chirpy]

The problem is that they can be false-positives, or an indication of file system corruption or a head-crash. With the latter, you only need one error to indicate a hosed disk. For peace of mind, I'd go with an OS restore on a new disk.

 

[DigiCrime]

peace of mind indeed, only way to fix the RPM database at this point anyway as i see no other way at the moment..