This project is mirrored from https://github.com/scylladb/scylladb.
Pull mirroring updated .
- Nov 18, 2015
-
-
Pekka Enberg authored
-
- Nov 17, 2015
-
-
Calle Wilund authored
If we get a partition with no row data, but statics, we should treat this as a row (include in count), but also make sure we skip to next partition if our page ends here. The "end partition" with zero rows but static data can also happen if we happen to resume paging by giving a column range exluding all data. In this case we should _not_ include it, since we have already provided the data in question in previous page. Fixes #556
-
Calle Wilund authored
1.) Should not reset to input query state if run repeatedly 2.) And if run repeatedly without input state, likewise keep the internal one active Fixes #560
-
Avi Kivity authored
"To keep compatibility with scylla-tools-java, it links /etc/scylla to /var/lib/scylla/conf. Problem on this patchset is, I added SCYLLA_HOME and SCYLLA_CONF on /etc/sysconfig/scylla-server. However, the file is marked as config file, it won't be automatically upgrade. If user doesn't upgrade the file manually, scylla-server still able to run with /var/lib/scylla/conf because we have symlink, but never switches to /etc/scylla."
-
- Nov 16, 2015
-
-
Takuya ASADA authored
-
Takuya ASADA authored
Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Takuya ASADA authored
Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Takuya ASADA authored
Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Takuya ASADA authored
Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Takuya ASADA authored
Fixes a problem on AMI which recently reported by Glauber Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Asias He authored
Skip sending the mutation if the cf is dropped after we call make_local_reader in stream_session::add_transfer_ranges(). Fix #550.
-
Paweł Dziepak authored
While the objects above max_manage_object_size aren't stored in the LSA segments they are still considered to be belonging to the LSA region and are evictable using that region evictor. Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
-
Avi Kivity authored
"This series adds the natural_endpoints API. It adds the implementation to the storage_service and to the storage_service API. After this series the noodtool command getendpoints should work. example: $ bin/nodetool getendpoints keyspace1 standard1 0x5032394c323239385030127.0.0.2 127.0.0.2"
-
Amnon Heiman authored
This patch adds the API for timeout messages and dropped messages. For dropped messages, origin has two APIs one for messages and one for command. droped messages return the number of messages per ver, so our API was rename to reflect that. For dropped messages (command) we currently do not have this logic of throwing messages before sending, so the API will always return 0. The total timeout API was removed and will be done on the jmx proxy level. Signed-off-by: Amnon Heiman <amnon@scylladb.com>
-
Gleb Natapov authored
Reviewed-by: Asias He <asias@scylladb.com>
-
Asias He authored
If listen_address is different than broadcast_address, we should use broadcast_address for the seeds list. Check and ask user to fix the configuration, e.g., $ scylla --rpc-address 127.0.0.1 --listen-address 127.0.0.1 --broadcast-address 192.168.1.100 --seed-provider-parameters seeds=127.0.0.1 Use broadcast_address instead of listen_address for seeds list: seeds={127.0.0.1}, listen_address=127.0.0.1, broadcast_address=192.168.1.100 Exiting on unhandled exception of type 'std::runtime_error': Use broadcast_address for seeds list
-
Asias He authored
-
Avi Kivity authored
"Fix various const correctness issues in gossiper code found while working on inter-node protocol."
-
Asias He authored
Abort accept and kill existing connections and wait for them. Fix tests in debug mode: ==29352==ERROR: AddressSanitizer: heap-use-after-free on address 0x60c000012758 at pc 0x00000211e371 bp 0x7ffe27369e10 sp 0x7ffe27369e00 READ of size 8 at 0x60c000012758 thread T0 #0 0x211e370 in transport::cql_server::do_accepts(int)::{ lambda(connected_socket, socket_address)#1}::operator()(connected_socket, socket_address)::{lambda(future<>)#1}::operator()(future) const (/home/asias/.dtest/dtest-AOBJua/test/node2/bin/scylla+0x211e370) #1 0x21a8090 in do_void_futurize_apply<transport::cql_server::do_accepts(int)::< lambda(connected_socket, socket_address)> mutable::<lambda(future<>)>, future<> > /home/asias/src/cloudius-systems/scylla/seastar/core/future.hh:1078 #2 0x217e861 in apply<transport::cql_server::do_accepts(int)::< lambda(connected_socket, socket_address)> mutable::<lambda(future<>)>, future<> > /home/asias/src/cloudius-systems/scylla/seastar/core/future.hh:1126 #3 0x223deb9 in _ZZN6futureIJEE12then_wrappedIZZN9transport10cql_server10do_accepts EiENUl16connected_socket14socket_addressE_clES4_S5_EUlS0_E_S0_EET0_OT_ENUlSA_E_ clI12future_stateIJEEEEDaSA_ (/home/asias/.dtest/dtest-AOBJua/test/node2/bin/scylla+0x223deb9)
-
Gleb Natapov authored
Write handler keeps track of all endpoints that not yet acked mutation verb. It uses broadcast address as an enpoint id, but if local address is different from broadcast address for local enpoints acknowledgements will come from different address, so socket address cannot be used as an acknowledgement source. Origin solves this by sending "from" in each message, it looks like an overhead, solve this by providing endpoint's broadcast address in rpc client_info and use that instead.
-
Tomasz Grabiec authored
Gleb Natapov (2): rpc: allow non const client_info to be passed to rpc handler rpc: allow auxiliary data to be stored in client_info
-
Pekka Enberg authored
The restart logic is wrong because C* had a bug in bf599fb5b062cbcc652da78b7d699e7a01b949ad and they fixed later and we translated the broken version. We must check if there is an existing endpoint state and call on_restart() hooks on that, not the newly available endpoint state. Spotted while inspecting the code. Acked-by: Asias He <asias@scylladb.com>
-
https://github.com/avikivity/scylla.gitTomasz Grabiec authored
Assorted patches that pave the way for native storage (while not committing us in any way).
-
https://github.com/avikivity/scylla.gitTomasz Grabiec authored
From Avi: Memtables do not use an allocating_section to guard against allocation failure, and hence can fail an allocation. Reproducible by changing perf_mutation to use an allocating type (bytes_type with a nontrivial size) and making the loop longer. Fix by using an allocating_section.
-
Avi Kivity authored
Without this, an allocation can fail, and we may not be able to reclaim memory.
-
Pekka Enberg authored
Signed-off-by: Pekka Enberg <penberg@scylladb.com>
-
Pekka Enberg authored
Signed-off-by: Pekka Enberg <penberg@scylladb.com>
-
Pekka Enberg authored
Signed-off-by: Pekka Enberg <penberg@scylladb.com>
-
- Nov 15, 2015
-
-
Avi Kivity authored
-
Tomasz Grabiec authored
-
Glauber Costa authored
Recently, I have introduced cf_stats into the database, propagating all the way back to the column family. The problem, however, is that some tests create a column family config themselves instead of going through make_column_family. That is ultimately ok if those tests are not expected to flush memtables. But if they are, the cf_stats pointer will be null and we will crash. Although there are many solutions to this, the one that is in tune with our current practices is to have the test that requires it provide an empty cf_stats storage area that can be written to. That's already how we handle the disk directory and other things like compaction properties. With this patch, test.py passes again. Signed-off-by: Glauber Costa <glommer@scylladb.com>
-
Glauber Costa authored
arbitrary 8G -> all_memory. Signed-off-by: Glauber Costa <glommer@scylladb.com>
-
Glauber Costa authored
This patch substitutes uint64_t for uint32_t as the type for commitlog_total_space_in_mb. Moving to 64 is not strictly needed, since even a signed 32-bit type would allow us to easily handle 2TB. But since we store that in the commitlog as a 64-bit value, let's match it. Moving from unsigned to signed, however, allow us to represent negative numbers. With that in place, we can change the semantics of the value slightly, so to allow a negative number to mean "all memory". The reason behind this, is that the default value "8GB", is an artifact of the JVM. We don't need that, and in many-shards configuration, each shard flushes the commitlog way too often, since 8GB / many_shards = small_number. 8GB also happens to be a popular heap size for C* in the JVM. For us, we would like to equate that (at least) with the amount of memory. The problem is how to do that without introducing new options or changing the semantics of existing options too radically. The proposed solution will allow us to still parse C* yaml files, since those will always have positive numbers, while introducing our own defaults. Signed-off-by: Glauber Costa <glommer@scylladb.com>
-
Avi Kivity authored
"Fixes part of #493 (few warnings still remains)"
-
Takuya ASADA authored
-
Takuya ASADA authored
-
Takuya ASADA authored
-
Takuya ASADA authored
-
Takuya ASADA authored
Signed-off-by: Takuya ASADA <syuu@scylladb.com>
-
Takuya ASADA authored
-