Daniel Jarboe wrote: >This usually happens around 7ish or noonish. I asked about the "linprox >kernel: end_request: I/O error, dev 5e:01 (dasd), sector " messages >before, but the last response I received was >"This is a known problem in the linux memory management code that only >shows up on s390. You can probably find a detailed explanation in the >mailing list archives. The good news is that the problem is completely >harmless, except for filling up the system logger." > >I was curious, because these messages seem to increase during failures, >and I am using the basevol+guestvol configuration. /dev/dasda1 is >shared r/o across several images. This is RH 7.2 with a rebuilt >2.5.STABLE3. Any ideas?
The I/O errors on a read-only device are most likely fixed by the latest service patch on developerWorks (for 2.4.7 check the August 2001 stream pages, patch linux-2.4.7-s390-14 as of 2003-07-07, problem ID 3144 "memory management: dirty bit clearing"). However, this is unlikely to have anything to do with the Squid crashes; note that the "User process fault" messages do not indicate a kernel problem in any way, but are simply a means to help debug user space application crashes. Bye, Ulrich -- Dr. Ulrich Weigand [EMAIL PROTECTED]