diff options
author | Matthias Radestock <matthias@rabbitmq.com> | 2014-03-04 11:13:00 +0000 |
---|---|---|
committer | Matthias Radestock <matthias@rabbitmq.com> | 2014-03-04 11:13:00 +0000 |
commit | f447541304502ed0f5b7f6f8085abd1446d6416f (patch) | |
tree | a775f0068fdc80134e44fa027b26433dc010f432 | |
parent | 8c1c0f4362f42d68e4654795c21c6507604da4c1 (diff) | |
download | rabbitmq-server-f447541304502ed0f5b7f6f8085abd1446d6416f.tar.gz |
increase IO_BATCH_SIZE
...so that beta->gamma conversion can keep up with alpha->beta
conversion.
Also correct comment - beta->gamma conversion can happen in a lot more
places than 'publish'
-rw-r--r-- | src/rabbit_variable_queue.erl | 14 |
1 files changed, 7 insertions, 7 deletions
diff --git a/src/rabbit_variable_queue.erl b/src/rabbit_variable_queue.erl index f6090d87..87576581 100644 --- a/src/rabbit_variable_queue.erl +++ b/src/rabbit_variable_queue.erl @@ -292,13 +292,13 @@ end_seq_id %% end_seq_id is exclusive }). -%% When we discover, on publish, that we should write some indices to -%% disk for some betas, the IO_BATCH_SIZE sets the number of betas -%% that we must be due to write indices for before we do any work at -%% all. This is both a minimum and a maximum - we don't write fewer -%% than IO_BATCH_SIZE indices out in one go, and we don't write more - -%% we can always come back on the next publish to do more. --define(IO_BATCH_SIZE, 1024). +%% When we discover that we should write some indices to disk for some +%% betas, the IO_BATCH_SIZE sets the number of betas that we must be +%% due to write indices for before we do any work at all. This is both +%% a minimum and a maximum - we don't write fewer than IO_BATCH_SIZE +%% indices out in one go, and we don't write more - we can always come +%% back on the next publish to do more. +-define(IO_BATCH_SIZE, 2048). %% next power-of-2 after ?CREDIT_DISC_BOUND -define(PERSISTENT_MSG_STORE, msg_store_persistent). -define(TRANSIENT_MSG_STORE, msg_store_transient). -define(QUEUE, lqueue). |