I have a CentOS 5.4 (RHEL 4.1.2-50 / Linux version 2.6.18-238.9.1.el5) system I remotely administer. It has a software RAID (mdadm v2.6.9) and recently had a hard drive replaced. The RAID rebuild successfully with no errors, and is showing no errors in the logs. However, a third party database technician tells me their program (using postgresql 8.3.11) has slowed down considerably after the RAID rebuilt.

I am trying to look at detailed logging from fsck, but since I am off-site and the disks being fsck'd are running the os, I have no other option then to run fsck in startup. I have touch'd /forcefsck and ran fsck with the -F switch to have it force on next boot, but have learned that fsck does has no log. I imagine the easiest way to do this is to create a fsck script that gets called before the software raid is mounted. I do not know to much about linux runlevels, as I mostly work with FreeBSD...

What is the best practice to getting a verbose output of a forced fsck saved to a file on startup? Thank you