Folks, We have an up-to date FreeBSD 5.4 Stable server that has suddenly started experiencing timeouts while reading and writing its 200Gb Western Digital SATA hard disk. We see at least one read / write timeout with kernel message every fifteen minutes or so and the box becomes sluggish while the IO is retried. The server has an Intel ICH5 SATA disk controller. I temporarily moved the disk to a recent 5.4 desktop machine equipped with the same chipset and the timeouts stopped, so it appears that the problem is not with the disk. I also tried Soren's ATA mkIII patches on the server, but the problem remains. Since the controller is built onto the Intel server board, we're tempted to try another brand of SATA disk. Does anyone have any ideas as to what could cause a previously happy server to start behaving this way? Regards, Tony. -- Tony Byrne
> I temporarily moved the disk to a recent 5.4 desktop machine equipped > with the same chipset and the timeouts stopped, so it appears that the > problem is not with the disk.So the problem exists on 5.S and not on 5.4-RELEASE? or is it vice versa?
Hello Robert, Saturday, May 28, 2005, 1:30:41 PM, you wrote:>> I temporarily moved the disk to a recent 5.4 desktop machine equipped >> with the same chipset and the timeouts stopped, so it appears that the >> problem is not with the disk.RB> So the problem exists on 5.S and not on 5.4-RELEASE? or is it vice versa? No. I should have said the the desktop is also a 5.S box. In fact both of the boxes have recent kernels and userlands from the stable branch (within last 7 days). Regards, Tony. -- Tony Byrne