A load-balancing kdb+ server

The script KxSystems/kdb/e/mserve.q can be used to start a load-balancing kdb+ server. The primary server starts a number of secondary servers (in the same host). Clients then send requests to the primary server which, transparently to the client, chooses a secondary server with low CPU load, and forwards the request there.

This set-up is useful for read operations, such as queries on historical databases. Each query is executed in one of the secondary threads, hence writes are not replicated.

Starting the primary server

The arguments are the number of secondary servers, and the name of a q script that will be executed by the secondary servers at start-up. Typically this script will read in a database from disk into memory.

$ q mserve.q -p 5001 2 startup.q

Client request

In the client, connect to the server with hopen

q)h: hopen `:localhost:5001

Synchronous messages are executed at the primary server.

q)h "xs: til 9"
q)h "xs"
0 1 2 3 4 5 6 7 8

Asynchronous messages are forwarded to one of the secondary servers, transparently to the client. The code below issues an asynchronous request, then blocks on the handle waiting for a result to be returned. This is called deferred synchronous.

q)(neg h) "select sym,price from trade where size > 50000" ; h[]

Deferred synchronous requests can also be made from non-q clients. For example, the example grid viewer code can be modified to issue a deferred synchronous request rather than a synchronous request by sending an async request and blocking on the handle in exactly the same way. The line

model.setFlip((c.Flip) c.k(query));

should be modified to

c.ks(query);
model.setFlip((c.Flip) c.k());

Source code

The complete source code for mserve.q is quite short, reproduced here for convenience.

/ start secondary servers
{value"\\q ",.z.x[1]," -p ",string x}each p:(value"\\p")+1+til"I"$.z.x 0;

/ unix (comment out for windows)
\sleep 1

/ connect to secondary servers
h:neg hopen each p;h@\:".z.pc:{exit 0}";h!:()

/ fields queries. assign query to least busy secondary server
.z.ps:{$[(w:neg .z.w)in key h;[h[w;0]x;h[w]:1_h w];                    /response
 [h[a?:min a:count each h],:w;a("{(neg .z.w)@[value;x;`error]}";x)]]}  /request