Hello! Is it possible that you are not closing your JDBC result sets (or corresponding Ignite QueryCursor)?
It's actually strange. The query in question seem rather bulky. It's not expected that it is sufficiently fast for you. Regards, -- Ilya Kasnacheev вт, 19 мая 2020 г. в 20:36, adipro <aditya.har...@zohocorp.com>: > My query is > > SELECT * FROM URLS WHERE APPNAME_ID = ? ORDER BY SCORE LIMIT ? > > That is giving me 0.7-1.0 msec read performance for a test run for about > some time. But the index with URLS (SCORE ASC, APPNAME_ID), I'm getting a > read performance of about 0.3 to 0.5 msec. i found it to be constant when > data grows.. whereas the index which you provided is getting delayed with > growing data. Ours is write intensive application. > > But the thing is in deployment servers we are constantly receiving these > warnings > > ``` > [22:26:31,046][WARNING][query-#48349][IgniteH2Indexing] Long running query > is finished [duration=3006ms, type=MAP, distributedJoin=false, > enforceJoinOrder=false, lazy=false, schema=PUBLIC, sql='SELECT > "__Z0"."ID" "__C0_0", > "__Z0"."URL" "__C0_1", > "__Z0"."SCORE" "__C0_2", > "__Z0"."APPNAME_ID" "__C0_3" > FROM "PUBLIC"."URLS" "__Z0" > WHERE "__Z0"."APPNAME_ID" = ?1 > ORDER BY 3 FETCH FIRST ?2 ROWS ONLY', plan=SELECT > __Z0.ID AS __C0_0, > __Z0.URL AS __C0_1, > __Z0.SCORE AS __C0_2, > __Z0.APPNAME_ID AS __C0_3 > FROM PUBLIC.URLS __Z0 > /* PUBLIC.IDX_2_URLS */ > /* scanCount: 1106428 */ > WHERE __Z0.APPNAME_ID = ?1 > ORDER BY 3 > FETCH FIRST ?2 ROWS ONLY > /* index sorted */, node=TcpDiscoveryNode > [id=9a62f525-98a8-43d6-85da-2270f1ee4e7a, > consistentId=9a62f525-98a8-43d6-85da-2270f1ee4e7a, addrs=ArrayList > [127.0.0.1, 172.20.42.17], sockAddrs=HashSet [/172.20.42.17:0, > /127.0.0.1:0], discPort=0, order=3, intOrder=3, > lastExchangeTime=1589893316075, loc=false, > ver=8.7.10#20191227-sha1:c481441d, isClient=true], reqId=262124, segment=0] > ``` > > > data metric at the point of warning -> > > ``` > [22:34:05,866][INFO][grid-timeout-worker-#407][IgniteKernal] > Metrics for local node (to disable set 'metricsLogFrequency' to 0) > ^-- Node [id=4f2b808c, uptime=04:19:01.321] > ^-- H/N/C [hosts=4, nodes=4, CPUs=192] > ^-- CPU [cur=0.03%, avg=13.79%, GC=0%] > ^-- PageMemory [pages=927396] > ^-- Heap [used=1521MB, free=62.86%, comm=4096MB] > ^-- Off-heap [used=3665MB, free=56.84%, comm=8392MB] > ^-- sysMemPlc region [used=0MB, free=99.98%, comm=100MB] > ^-- default region [used=3665MB, free=55.26%, comm=8192MB] > ^-- metastoreMemPlc region [used=0MB, free=99.94%, comm=0MB] > ^-- TxLog region [used=0MB, free=100%, comm=100MB] > ^-- Ignite persistence [used=3622MB] > ^-- sysMemPlc region [used=0MB] > ^-- default region [used=3622MB] > ^-- metastoreMemPlc region [used=0MB] > ^-- TxLog region [used=0MB] > ^-- Outbound messages queue [size=0] > ^-- Public thread pool [active=0, idle=0, qSize=0] > ^-- System thread pool [active=0, idle=200, qSize=0] > ^-- Striped thread pool [active=2, idle=198, qSize=0] > ``` > > Can you please tell why these warnings are coming. Although in client side, > I added a check if query delays to print a warning. But it didn't through > any warnings in client machine. It's weird why this warning is coming in > server logs. > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >
