![]() | |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
Well, I figured out the problem, and was able to get the code to successfully complete. It turns out a default transaction log file of only 256MB was not large enough. I recreated it at 4096MB, and everything work okay. |
#3
| |||
| |||
|
|
Well, I figured out the problem, and was able to get the code to successfully complete. It turns out a default transaction log file of only 256MB was not large enough. I recreated it at 4096MB, and everything work okay. |
#4
| |||
| |||
|
|
Well, I figured out the problem, and was able to get the code to successfully complete. It turns out a default transaction log file of only 256MB was not large enough. I recreated it at 4096MB, and everything work okay. |
#5
| |||
| |||
|
|
Also try doing the load after issuing the statement: set session with on_logfull=commit; |
#6
| |||
| |||
|
|
Ingres Forums wrote: Well, I figured out the problem, and was able to get the code to successfully complete. *It turns out a default transaction log file of only 256MB was not large enough. *I recreated it at 4096MB, and everything work okay. Is the reason that you 'needed' a larger transaction log file, that you were attempting to run the operation as one transaction? I assume that there is no business requirement to load all 1 million Customer entities and 1 million PricingMethod entities in one or two transactions. Try to keep your transactions as small as the business limitations allow. If you 'commit' after groups of 100 or even 1000, then you would probably not need a 4Gb transaction log file, but would probably have completed with the default tx-log. When copying data, it is tempting to just bung in the whole lot in one go, but that is resource demanding. Additionally, for every transaction log file, no matter what its size, I can devise a data load that will cause the failure that you experienced. Remember the method for eating chocolate elephants. *- A little bit at a time. Michael |
#7
| |||
| |||
|
|
Also what about the round trips to the database involved? Is a command executed once for each row? I guess this could add to the overhead too |
![]() |
| Thread Tools | |
| Display Modes | |
| |