OK, after the huge emergency, "You can't attend the last day of open world because our problems are too big." After my debugging efforts, how did the problem of the slow writes to the redo log files resolve?
1. Management asked that the entire database be moved to the NAS disks because they initially seemed faster. I had to let them know, that, no, the writes there are not faster.
2. The efforts to cross the great divide and get the storage manager in the other company to actually look at his configuration resulted in, "Since the writes are slow on two different pieces of our machinery, it can't be our fault - it must be oracle. I'm debugging nothing."
3. The other company informed us, "By the way, we will be installing new hardware in a couple of weeks at the same time that you are making a major application upgrade." How does that sound for a prospect of a smooth transition? This was followed by an email, "Claudia, are you working with the other company on this?" Not only am I not working with them, I never heard of it!
4. Management informed me, "Since there will be a hardware change soon, don't bother to follow up on this problem."
So they will install the hardware, we will deploy the new application version, and there will be storm and drama about the excessive waits for redo log writes. I think that I should change my name to Cassandra.
A few basic notes on Oracle Database Administration.
Saturday, September 27, 2008
- ▼ September (5)