|
HP OpenVMS Systemsask the wizard |
|
The Question is: We use DECBasic and RMS files for our applications. At one of our 100+ sites, they sporadically receive "Error 12 - ?Fatal system I/O failure" when issuing an OPEN or GET command. This occurs across multiple disk spindles and with multiple application executables -- so my first thought would be a storgae works or controller problem. However, there are no VMS generated errors. The problem manifests itself for anywhere from 1-2 seconds to 10 minutes. It usually occurs when the system is under a heavy load, but no always. It can go away throuigh no external action for several days, or it can occur 1200 times a day for several d ays. On a typical day it will occur 1-3 times. We are in the process of replacing the single HSJ40 controller with dual HSJ52s -- but don't feel comfortable that this will correct the problem. Any ideas? The Answer is : Acquire and apply the current ECOs for OpenVMS -- an upgrade to OpenVMS Alpha V7.1-2 can be an easy way to perform this -- and check the firmware revision in the host PALcode, I/O controller(s), and storage controller(s). Upgrade to current. Also check for hardware errors at the site, in such subsystems as the CPU, memory, and I/O. Check for any additional information available -- OpenVMS will generally have RMS information available in the STS and STV cells. In various BASIC dialects on OpenVMS, see the RMSSTATUS and/or VMSSTATUS calls -- these values can provide you with additional information on the failure. Also check the values of the available process quotas on the systems, paying particular attention to potential quota starvations. If the problem persists, please contact the Compaq Customer Support Center for assistance.
|