query causes signal 11 error and Vertica(10) going down
sreeblr
- Select Field - Employee
2021-03-31 09:36:02.780 Init Session:0x7f0e2e0a4700-a000000171acc5 [Txn] Begin Txn: a000000171acc5 'with TOPTEN ESDS ( ESDsCount, ESDsName) as ( select count( ESD_TYPE_NAME) as ESDsCount,PFE_ ESD_NAME as ESDsName from ESD_GRP_RESULTS GRP_RES join (select ADD_MONTHS(max(cast(RPX_PERIOD_END_DT as date)), -36) as RPX_PERIOD_START_DT, max(cast(RPX_PERIOD_END_DT as date)) as RPX_PERIOD_END_DT from ESD_GRP_RESULTS) AS GRP_RES_1 on cast(GRP_RES.RPX_PERIOD_END_DT as date) = GRP_RES_1.RPX_PERIOD_END_DT and cast(GRP_RES.RPX_PERIOD_START_DT as date) = GRP_RES_1.RPX_PERIOD_START_DT group by @ENTITY, PFE_ ESD_NAME ) select sum(TOPTEN ESDS. ESDsCount) as ESDsCount,TOPTEN ESDS. ESDsName from TOPTEN ESDS group by EsdName order by EsdCount desc limit 10'
2021-03-31 09:34:33.183 Init Session:0x7fc67dfca700-a00000008f6142 [Main] Handling signal: 11
2021-03-31 09:34:34.225 Init Session:0x7fc67dfca700-a00000008f6142 [Main] Received fatal signal SIGSEGV.
2021-03-31 09:34:34.225 Init Session:0x7fc67dfca700-a00000008f6142 [Main] Info: si_code: 1, si_pid: 0, si_uid: 0, si_addr: (nil)
SQL was very simple -not sure why it would cause node to go down. i am on vertica 10.0.1-0 and red hat 7.9
i saw old 7.2.x link https://forum.vertica.com/discussion/237834/how-to-investigate-a-panic not sure if this needs a 10.x patch?
Answers
@sreeblr - Can you please post the DDL for the ESD_GRP_RESULTS table?
@sreeblr Can you also check the ErrorReport.txt and search for transition a000000171acc5. The reason I am asking to check the ErrorReport to make sure the query you have mentioned is causing the node crash or something else.
Just curious Is this a new query or just fail or cause to bring the node first time?
yes @Nimmi_gupta found a000000171acc5 in ErrorReport.txt
Can you share the snippet of the ErrorReport. Backtrace from the Errorreport will show at what stage (function call) query is failing.
Did you try to run the same query again and did it crash the node again. just to make sure this query is a real problematic query and we can focus on the query and collect the corresponding information.
opened ticket SD02870840. they have Jira VER-76815 opened for this issue
Perfect !!