When we call merge_keyspaces, we end up calling db.add_keyspace, which means we
don't do parts of creation of a keyspace - like creating the directory. This works
in some situations where all the work was already done, but not in others.
We should be calling create_keyspace instead, which will take care of all that
for us. That will include creating a directory when one is needed.
Signed-off-by: Glauber Costa <glommer@cloudius-systems.com>
There is only one user in tree, and as Tomek pointed out, it is buggy. The
reason is that ksm is a shard-local structure, and it is currently used
indiscriminately by all shards which can easily lead to problems.
We could fix it with some tricks, but it is way better and safer to make sure
the callers are doing it right instead. There is only one caller, so let's fix
that.
Signed-off-by: Glauber Costa <glommer@cloudius-systems.com>
The two separate functions can now be merged. As a result, the code
that generates statistics data is now much easier to understand.
Signed-off-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Assert should be completely avoided. Instead, we should trigger an
exception, allowing the db to proceed with a "smooth" shutdown.
Signed-off-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
"This is the current patchset to flush and persist schema changes to disk.
It is not perfect, in the sense that older changes still in flight won't be
waited for. But as we discussed - at this moment we'll just note that, and
leave the fix for later"
KW_MINCOMPACTIONTHRESHOLD and KW_MAXCOMPACTIONTHRESHOLD shouldn't be
allowed, they are properties of compaction strategy not the table itself.
KW_COMPACTION and KW_COMPRESSION should be allowed.
Reported-by: Shlomi Livne <shlomi@cloudius-systems.com>
Signed-off-by: Paweł Dziepak <pdziepak@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Apparently, after writing a new sstable, with write_components(), it
is necessary to load() it. I'm not sure why, but we get a crash on
an aio to a closed file descriptor if we don't.
Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Instead of requiring the user to subclass a "sstable_creator" class to
specify how to create a new sstable (or in the future, several of them),
switch to an std::function.
In practice, it is much easier to specify a lambda than a class, especialy
since C++11 made it easy to capture variables into lambdas - but not into
local classes.
The "commit()" function is also unnecessary. Then intention there was to
provide a function to "commit" the new sstables (i.e., rename them).
But the caller doesn't need to supply this function - it can just wait
for the future of the end of compaction, and do his own committing code
right then.
Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
"This series adds the cache_service definition file with a stub implementation.
The API is based on the CacheServiceMBean interface.
The stubbed API return stubbed value of the correct type."
This adds a stub implementation for the cache_service. It also register
the API doc.
The cache_service doc will be available at:
http://localhost:10000/api-doc/cache_service/
The stub implementation returns the correct value type with stubbed
values.
After this patch the following path will be available:
/cache_service/row_cache_save_period
/cache_service/key_cache_save_period
/cache_service/counter_cache_save_period
/cache_service/row_cache_keys_to_save
/cache_service/key_cache_keys_to_save
/cache_service/counter_cache_keys_to_save
/cache_service/invalidate_key_cache
/cache_service/invalidate_counter_cache
/cache_service/row_cache_capacity
/cache_service/key_cache_capacity
/cache_service/counter_cache_capacity
/cache_service/save_caches
Signed-off-by: Amnon Heiman <amnon@cloudius-systems.com>
This adds the definition file for the cache service, the API is based on
the CacheServiceMBean definition and contain the following command:
get_row_cache_save_period_in_seconds
set_row_cache_save_period_in_seconds
get_key_cache_save_period_in_seconds
set_key_cache_save_period_in_seconds
get_counter_cache_save_period_in_seconds
set_counter_cache_save_period_in_seconds
get_row_cache_keys_to_save
set_row_cache_keys_to_save
get_key_cache_keys_to_save
set_key_cache_keys_to_save
get_counter_cache_keys_to_save
set_counter_cache_keys_to_save
invalidate_key_cache
invalidate_counter_cache
set_row_cache_capacity_in_mb
set_key_cache_capacity_in_mb
set_counter_cache_capacity_in_mb
save_caches
Signed-off-by: Amnon Heiman <amnon@cloudius-systems.com>
If local mutation write takes longer then write timeout mutation will
be deleted while it is processed by database engine. Fix this by storing
mutation in shared pointer and hold to the pointer until mutation is
locally processed.
