batchstatement
TL/DR:
- only 1 out of 14 statements in a batch do not mutate the partition..
- no error is logged in the application layer, or Cassandra system.log or
Cassandra debug.log
- using C#, and Datastax latest driver
- cluster is a 1-node, dev setup.
- datastax driver configured with LOCAL_QUORUM at the session, and
statement level.
- using preparedstatements.. 1,000% sure there's no typo.. (but I've been
wrong before)
I have about 14 statements that get batched up together. They're updating
at most 2, maybe 3 denormalized tables.. all the same user object, just
different lookup keys.
To help visualize, the tables look a little like these.. abbreviated..
User table..
CREATE TABLE user (u_id uuid, act_id uuid, ext_id text, dt_created
timeuuid, dt_mod timeuuid, is_group Boolean, first_name text)
Users By Account (or plan)
CREATE TABLE user_by_act (act_id uuid, u_id uuid, first_name text)
User By external identifier
CREATE TABLE user_by_ext (ext_id text, u_id uuid, act_id uuid, first_name
text)
I create a batch that updates all the tables.. various updates are broken
out into separate statements, so for example, there's a statement that
updates the external ID in the 'user' table.
UPDATE user_by_ext SET ext_id = :ext_id WHERE u_id = :u_id
This particular batch has 14 statements total, across all 3 tables. They
are only updating at most 3 partitions.. a single partition may have 4 or
more statements to update various parts of the partition. e.g. first name
and last name are a single statement added to the batch.
Here's the problem... of those 14 statements.. across the 3 partitions...
ONE and ONLY ONE update doesn't work.. Absolutely every other discreet
update in the whole batch works.
List boundStatements = new
List();
// *
// user table
boundStatements.Add(SessionManager.UserInsertStatement.Bind(new
{ u_id= user.UserId, act_id = user.ActId, dt_created = nowId, dt_mod =
nowId, is_everyone = user.IsEveryone, is_group = user.IsGroup }));
if (!string.IsNullOrWhiteSpace(user.ExtId))
//
// this statement gets added to the list.. it is part of the batch
// but it NEVER updates the actual field in the databse.
// I have moved it around.. up, down... the only thing that works
// is if I call execute on the first binding above, and then add the rest
// of these as a separate batch.
boundStatements.Add(SessionManager.UserUpdateExtIdStatement.Bind(new { u_id
= user.UserId, ext_id = user.ExtId, dt_mod = nowId }));
//
if (!string.IsNullOrWhiteSpace(user.Email))
boundStatements.Add(SessionManager.UserUpdateEmailStatement.Bind(new { u_id
= user.UserId, email = user.Email, dt_mod = nowId }));
BoundStatement userProfile =
CreateUserProfileBoundStatement(nowId, user);
if (userProfile != null)
boundStatements.Add(userProfile);
// *
// user_by_act table
CreateUserAccountInsertBoundStatements(boundStatements, user,
nowId);
// *
// user_by_ext table
if (!string.IsNullOrWhiteSpace(user.ExtId))
{
boundStatements.Add(SessionManager.UserExtInsertStatement.Bind(new { ext_id
= user.ExtId, act_id = user.ActId, dt_created = nowId, dt_mod = nowId,
is_group = user.IsGroup, u_id = user.UserId }));
BoundStatement userByExtProfile =
CreateUserByExtProfileBoundStatement(nowId, user);
if (userByExtProfile != null)
boundStatements.Add(userByExtProfile);
if (!string.IsNullOrWhiteSpace(user.Email))
boundStatements.Add(SessionManager.UserExtUpdateEmailStatement.Bind(new {
ext_id = user.ExtId, email = user.Email, dt_mod = nowId }));
}
--
Randy Lynn
rl...@getavail.com
office:
859.963.1616 <+1-859-963-1616> ext 202
163 East Main Street - Lexington, KY 40507 - USA
Re: Best approach for node decommission
If it helps you can use 'nodetool netstats' to monitor . If you do not see any streams stuck there , then you should be good . Also if you are concerned , you can run "nodetool repair" after the decommission just to make sure . I think what you are doing is okay . If the node is alive then nodetool decommission is best approach . If the node is dead then you have to use nodetool removenode On Thu, Jul 12, 2018 at 10:45 AM, rajasekhar kommineni wrote: > Hi All, > > Can anybody let me know best approach for decommissiong a node in the > cluster. My cluster is using vnodes, is there any way to verify all the > data of the decommissioning node has been moved to remaining nodes, before > completely shutting down the server. > > I followed below procedure : > > 1) nodetool flush > 2) nodetool repair > 3) nodetool decommission > > The aggregate of Load before node3 decommission is 1411.47 and after is > 1380.15. Can I ignore the size difference and treat all the data of node3 > has been moved to other nodes. > > I am looking for good data validation process with out depending on > Application team for verification. > > *Total load : 1411.47* > > – Address Load Tokens Owns Host ID Rack > UN node1 220.48 MiB 256 ? ff09b08b-29c1-4365-a3b7-1eea51f7d575 rack1 > UN node2 216.53 MiB 256 ? 4b565a31-4c77-418f-a47f-5e0eb2ec5624 rack1 > UN node3 64.52 MiB 256 ? 12b29812-cc60-456c-95a9-0e339c249bc8 rack1 > UN node4 195.84 MiB 256 ? 0424a882-de4f-4e6a-b642-6ce9f4621e04 rack1 > UN node5 179.07 MiB 256 ? 2f291a2e-b10d-4364-8192-13e107a9c322 rack1 > UN node6 213.75 MiB 256 ? cf10166b-cfae-44fd-8bca-f55a4f9ef491 rack1 > UN node7 158.54 MiB 256 ? ef8454c7-3005-487a-a3d4-e0065edfd99f rack1 > UN node8 162.74 MiB 256 ? 7d786e46-1c11-485c-a943-bbcca6729ae1 rack1 > > *Total Load : 1380.15* > > – Address Load Tokens Owns Host ID Rack > UN node1 229.04 MiB 256 ? ff09b08b-29c1-4365-a3b7-1eea51f7d575 rack1 > UN node2 225.52 MiB 256 ? 4b565a31-4c77-418f-a47f-5e0eb2ec5624 rack1 > UN node4 195.84 MiB 256 ? 0424a882-de4f-4e6a-b642-6ce9f4621e04 rack1 > UN node5 179.07 MiB 256 ? 2f291a2e-b10d-4364-8192-13e107a9c322 rack1 > UN node6 229.4 MiB 256 ? cf10166b-cfae-44fd-8bca-f55a4f9ef491 rack1 > UN node7 158.54 MiB 256 ? ef8454c7-3005-487a-a3d4-e0065edfd99f rack1 > UN node8 162.74 MiB 256 ? 7d786e46-1c11-485c-a943-bbcca6729ae1 rack1 > > Thanks, > >
Re: cassandra cluser sizing
Jeff, thank you very much for reply. Will try to use 4TB per instance. If I understand it correctly level compaction can lead to 50% https://docs.datastax.com/en/dse-planning/doc/planning/planningHardware.html Regarding the question of running multiple instances per server, am I correct that in case of 3.11 instances and having several disks dedicated for each instance, running multiple instances per server is ok? On Thu, Jul 12, 2018 at 5:47 PM Jeff Jirsa wrote: > > You can certainly go higher than a terabyte - 4 or so is common, Ive heard of > people doing up to 12 tb with the awareness that time to replace scales with > size on disk, so a very large host will take longer to rebuild than a small > host > > The 50% free guidance only applies to size tiered compaction, and given your > throughput you may prefer leveled compaction anyway. With leveled you should > target 30% free for compaction and repair > > You don’t need more than one Cassandra instance per host for 4tb but you may > want to consider it for more than that - multiple instances are especially > useful if you have multiple (lots of) disks and are running Cassandra before > CASSANDRA-6696 (which made jbod safer). > > -- > Jeff Jirsa > > > > On Jul 12, 2018, at 7:37 AM, Vitaliy Semochkin wrote: > > > > Hi, > > > > Which amount of data Cassandra 3 server in a cluster can serve at max? > > The documentation says it is only 1TB. > > If the load is not high (only about 100 requests per second with 1kb > > of data each) is it safe to go above 1TB size (let's say 5TB per > > server)? > > What will be safe maximum disk size a server in such cluster can serve? > > > > Documentation also says that compaction requires to have %50 of disk > > occupied space. In case I don't have update operations (only insert) > > do I need that much extra space for compaction? > > > > In articles (outside Datastax docs) I read that it is a common > > practice to launch more than one Cassandra server on one physical > > server in order to be able use more than 1TB of hard driver per > > server, is it recommended? > > > > - > > To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org > > For additional commands, e-mail: user-h...@cassandra.apache.org > > > > - > To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org > For additional commands, e-mail: user-h...@cassandra.apache.org > - To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org For additional commands, e-mail: user-h...@cassandra.apache.org
