Gwyneth Llewelyn
|
24b5ac01bc
Started revamping information on the README
|
7 years ago |
Gwyneth Llewelyn
|
3802c5d082
Added information that leveldb is a possible flag for database type.
|
7 years ago |
Gwyneth Llewelyn
|
c3bd1d79d1
Fix some bugs regarding leveldb.
|
7 years ago |
Gwyneth Llewelyn
|
a24152d11e
Simple mmap testing suite.
|
7 years ago |
Gwyneth Llewelyn
|
94261df406
Testing with leveldb
|
7 years ago |
Gwyneth Llewelyn
|
b95711d334
Testing with levels
|
7 years ago |
Gwyneth Llewelyn
|
e1d1677430
Reverting ‘no memory’ flag to use memory-mapping again
|
7 years ago |
Gwyneth Llewelyn
|
0ae1302925
Memory tests. go test -v will show mmap being used on files of increased RAM size, in multiples of 128 Mbytes.
|
7 years ago |
Gwyneth Llewelyn
|
ccf65e4714
Do some memory tests. Use go test -run ‘’ to experiment with some settings
|
7 years ago |
Gwyneth Llewelyn
|
57a9a2cd9b
Attempt to consume as little memory as possible, tweaking badger options.
|
7 years ago |
Gwyneth Llewelyn
|
1c3065d35d
Attempting to make the search iterators faster…
|
7 years ago |
Gwyneth Llewelyn
|
711f70ee37
Upgrade to use an external configuration file; preparation to allow several different database backends, since I’m having trouble with keeping memory limits; also, badger and bolt seem to be way too slow.
|
7 years ago |
Gwyneth Llewelyn
|
add89f31f9
Configuration now is read from file (using the Viper library). Some changes need to be made to support other databases.
|
7 years ago |
Gwyneth Llewelyn
|
16bd60461c
Added logging options to config file as well.
|
7 years ago |
Gwyneth Llewelyn
|
373224b041
Because this is getting increasingly complicated, there is now a configuration file.
|
7 years ago |
Gwyneth Llewelyn
|
01ac1dbb52
Attempt to use BuntDB instead of Badger
|
7 years ago |
Gwyneth Llewelyn
|
663c41bd0f
Reduce to batches of 100,000 keys
|
7 years ago |
Gwyneth Llewelyn
|
37808f31a7
Attempting to change all code to work with the new version of Badger (changed everything!)
|
7 years ago |
Gwyneth Llewelyn
|
01a083b88a
More (failed!) attempts to limit memory consumption by reducing the ValueLogFileSize
|
7 years ago |
Gwyneth Llewelyn
|
908ac9477e
While memory errors seem to be fixed, there is a strange error related to opening the correct directory under FastCGI.
|
7 years ago |
Gwyneth Llewelyn
|
1a196640ff
Changing to use BatchSetAsync; attempting to deal with excessive memory issues.
|
7 years ago |
Gwyneth Llewelyn
|
e901145ce8
Changed BATCH_BLOCK back to 100,000.
|
7 years ago |
Gwyneth Llewelyn
|
2ae117743a
Added comment regarding the new BatchSet import.
|
7 years ago |
Gwyneth Llewelyn
|
446bed38fc
Using goroutines with BatchSet to speed up the process of importing.
|
7 years ago |
Gwyneth Llewelyn
|
140ba84e40
Attempt to use the new interface for badger
|
7 years ago |
Gwyneth Llewelyn
|
bf56c63448
Deal with the changes of the badger API
|
7 years ago |
Gwyneth Llewelyn
|
cc83f5d69e
On the interactive shell, allows searches by either avatar name or UUID. UUID is just a special case of a 36-character string with a valid format, which is checked with a regex.
|
7 years ago |
Gwyneth Llewelyn
|
d4a4a36251
Adds -nomemory switch to try to reduce memory consumption. This is the default for FastCGI.
|
7 years ago |
Gwyneth Llewelyn
|
accad7a5e8
Added support for keeping the grid of origin as part of the database entries. This required some JSON marshalling. Added some stopwatches, too, to give an idea how long it takes to include W-Hat’s database.
|
7 years ago |
Gwyneth Llewelyn
|
b1dbdd9015
Added more information about importing databases.
|
7 years ago |