Hi all, I recently built a new news server using RHL 5.0. My news spool consists of several brand new SCSI drives striped together in a raid0 setup. The SCSI controller is a Buslogic. The system is also using a PCI IDE controller and drives to hold the OS, over view database, history and active files, etc. All in all this system is VERY fast, taking in multiple full feeds and rarely using more than 5% CPU and only about a quarter of the 256 megs of RAM (except during expire of course). Anyway, it only takes a day or two before the file system on the striped partition starts to become corrupted. I start seeing directories that look like this: ?r-S--srwt 3855 3855 3855 252645135 Jan 2 1978 recreational Or directories that look ok, but if I go into one and do 'ls' I get errors like this: ls: .: No such file or directory I've completely reformatted the partition but it stills winds up getting corrupted. Also, during the nightly expiration process I often get a message like this: Warning - running *really* low on DMA buffers. Is this related? Any ideas what's happening here? Are there any known problems using md under RHL 5.0? Thanks in advance, Mark -- PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES! http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe" as the Subject.
