bg0l+@andrew.cmu.edu (Bruce E. Golightly) (01/25/89)
We are in the process of testing Ingres under VMS 5.0. So far, nothing really strange has happened, but we're not quite finished. We handle such testing on a VAX 11/750 to check for unpleasant side effects before moving the new version into out multi-node VAXCluster, since the cluster is used for production work that we can't afford to have break (like my paycheck comes from there!). We have beta of Ingres 6, but won't be moving to production for a while. We have applications up using STAR, which presently isn't available for 6.1. When we get that, we'll run through the same testing process for it, too.
davek@rtech.rtech.com (Dave Kellogg) (01/30/89)
In article <gXr-DNy00WB8IZ8GUO@andrew.cmu.edu> bg0l+@andrew.cmu.edu (Bruce E. Golightly) writes: >We are in the process of testing Ingres under VMS 5.0. So far, nothing >really strange has happened, but we're not quite finished. We handle such >testing on a VAX 11/750 to check for unpleasant side effects before moving >the new version into out multi-node VAXCluster, since the cluster is used >for production work that we can't afford to have break (like my paycheck >comes from there!). I coordinated the initial release of INGRES 5 on VAX/VMS V5, so I'm pretty close to this topic. Internally, very very little changed in INGRES. We had to re-macro and re-link one shareable image (SYS_INGRES:[INGRES.BIN]- IISYSSERV.EXE) and change the build script to accept VMS V5 as a valid version. Checkenv, also, needed to be updated to accept V5 as a valid version. Because INGRES uses the VMS lock manager we didn't have to worry about synchronization problems or re-writing device drivers to understand spinlocks. I did the initial on-site testing at a customer site in July, and since then I know of no problems specific to or caused by VMS V5, except for a simple bug which incorrectly issues a warning in the build procedure. If you're particularly nervous you might call technical support (have your contact person call them, that is) to verify what I'm saying. (Sorry about the warning message, it was my fault..) Thanks, David Kellogg
bg0l+@andrew.cmu.edu (Bruce E. Golightly) (02/01/89)
So far, so good, as they say. Thanks for the reassurance, Dave. I probably should point out that we're checking things other than Ingres for problems. We have a lot of other stuff that has to be checked out and running properly before we can change over the cluster. We've got COBOL, C, DCL and thirty-seven billion other odds and ends that all need to function perfectly when we cut over. Testing should be through in another week or so. I'll post a follow up later. Bruce. P.S. Dave, watch out for the pool when your co-workers are around. See you in New Orleans?