![]() |
|
|
|
|
|
|
2
25th April 10:07
External User
Posts: 1
|
Hi Manny,
Are all inserts slow or just this one? Are you using RAID 5? How much data is "some data" that is being inserted? A few rows, a few hundred wide rows, or a few thousand rows? You could set log_trace and see how much logging is being done. This would help show if there were many pages being written, or if there were log waits. You might also want to look at DM420 tracepoint data. A temporary table will reside in the DMF cache while it can, but will force pages to be written to disk if the cache is not sized large enough. Just for grins you could run a lock_trace and see if it shows anything. It's unlikely, but you never know. There are many, many things that could cause this type of issue. One of the first goals is to determine if the issue is systemic, or isolated to a single process or event. From there just keep logically digging until you find your answer. BTW, we have a really good Ingres Performance Tuning course that covers these and many other useful topics! :-) See http://www.comp-soln.com/ingres.html for more information. Best wishes. Chip Nickolett (ChipN@Comp-Soln.com) US - Comprehensive Solutions www.Comp-Soln.com UK - Comprehensive Solutions International, Ltd. www.Comp-Soln.co.uk |
|
|
4
27th April 14:27
External User
Posts: 1
|
Hi Manny,
Are all inserts slow or just this one? Are you using RAID 5? How much data is "some data" that is being inserted? A few rows, a few hundred wide rows, or a few thousand rows? You could set log_trace and see how much logging is being done. This would help show if there were many pages being written, or if there were log waits. You might also want to look at DM420 tracepoint data. A temporary table will reside in the DMF cache while it can, but will force pages to be written to disk if the cache is not sized large enough. Just for grins you could run a lock_trace and see if it shows anything. It's unlikely, but you never know. There are many, many things that could cause this type of issue. One of the first goals is to determine if the issue is systemic, or isolated to a single process or event. From there just keep logically digging until you find your answer. BTW, we have a really good Ingres Performance Tuning course that covers these and many other useful topics! :-) See http://www.comp-soln.com/ingres.html for more information. Best wishes. Chip Nickolett (ChipN@Comp-Soln.com) US - Comprehensive Solutions www.Comp-Soln.com UK - Comprehensive Solutions International, Ltd. www.Comp-Soln.co.uk |
|