Various Questions

A few questions for anyone that might know: 1) Update Statistics a) What sort of regularity would you suggest for updating statistics on a production system, with lots of activity, with at least 22 hours a day of active time? We have one customer using a Pyramid MIS-S server, Online 4.1, and ESQL/C, 4GL, and ISQL, with a 3 gigabyte mirrored database. Last time I updated stats, it took 45 minutes, so its probably not reasonable to run this on a regular basis. We have another customer with soon to be the same setup, but a 60 gig database. How often should we update statistics there, and how long would you expect it to take? b) Is it safe to abort an update statistics? Maybe my memory is wrong, but I seem to remember in the earlier versions of Informix, that running update statistics was actually considered dangerous. Anyone remember this? In general, was updating statistics ever a problem and is it safe to abort if it takes longer than we can stand? c) Why is it necessary to update statistics, when isql table status shows the correct number of rows? 2) Installing 4.1 Online on a Pyramid 9845 a) We are currently running 4.0 Online on our Pyramid 9845, with Osx5.0. We have been told that 4.1 cannot be installed unless we upgrade the the operating system to Osx5.1. Can someone please explain to me the details of why this is true. 3) Fast Indexing a) We hear there is a possibly undocumented feature to created indices faster than normal. The environment variables SORT_INDEX 3, PSORT, and PSORT_NPROCS 0 have been mentioned, but I cant find anything explaining this in any detail. I really dont like the idea of stuffing something like this in production until I understand more of what it means, but I think we want/need it. I think this is 5.0 added feature, but I have been told that it can run on a 4.1 system, but was just undocumented. Is it possible to run under earlier releases, and can you explain the procedure and background? We would love to install 5.0 on all our platforms, but it is not yet available. 4) Create Index and Locks a) We have one customer with a table that has over a million rows, and 3 keys on a table. The first time we tried to add a key to the loaded table, we overflowed the lock table, and did great damage. We were told (by Informix) a lock is used for each row TIMES each key, so 3 million locks would have been necessary (which you cant configure). If we did and exclusive lock on the table, we overflowed the logs. Stopping logging meant taking the entire application down, and that was unreasonable for any significant length of time. An exclusive database lock has the same problem. To bypass the problem, we created an empty table with keys intact, and transferred the rows to the new table. It took a little longer than forever, but it did eventually work. What would you suggest as the procedure for adding keys to tables that have over a million rows? I know, this is a *lot* of questions from someone not from New Jersey, but we have our hands full here, and are trying desperately to avert disaster. Naomi -- Naomi Walker (aka N7FSA) naomi%anasaz.UUCP@asuvax.eas.asu.edu Enthusiasm is caught, What if, there were no hypothetical not taught. situations.......