Options
[20032] [20032] Operation canceled.
avoronin
✭
I have this error for no apparent reason [20032] [20032] Operation canceled.
Certain select with a few joins is executed by SSIS via ADO NET driver to Vertica 1 node server. Result set up to 700000 rows was successfully retrieved dozens of times before that taking up to 90 seconds. The query was single query on the server at that time. Then suddenly I get " [20032] [20032] Operation canceled.". Running this query manully produces no problems -- first 200 rows gotten in 20 seconds.
What might have happen? And where can I check for deeper error info?
Maybe some timeout should be increased?
0
Answers
Query plan -- runtime 23 seconds
+-SORT [Cost: 1M, Rows: 22M (NO STATISTICS)] (PATH ID: 1)
Order: A.POSTINGDATE ASC, A.HASH ASC
+---> JOIN (INTERPOLATED) [LeftOuter] [Cost: 563K, Rows: 22M (NO STATISTICS)] (PATH ID: 3) Inner (SORT ON JOIN KEY)
Join Cond: (a0.A_SD_LINES_ID = t2.A_SD_LINES_ID) AND (a0.Actual_date = t2.actual_date)Materialize at Input: t2.A_SD_LINES_ID, t2.sd_POSTINGDATE_A_Date_ID, t2.actual_dateFilter: (a2.Key >= 20190101)Filter: (a2.Key <= 20190101)
+-- Outer -> JOIN (INTERPOLATED) [LeftOuter] [Cost: 279K, Rows: 22M (NO STATISTICS)] (PATH ID: 4)
Join Cond: (a0.A_SD_LINES_ID = a1.A_SD_LINES_ID) AND (a0.Actual_date = a1.actual_date)
+---> STORAGE ACCESS for a0 [Cost: 102K, Rows: 22M] (PATH ID: 6)
Projection: A.A_SD_LINES_superMaterialize: a0.A_SD_LINES_ID, a0.SD_LINES_ID
+-- Inner -> STORAGE ACCESS for a1 [Cost: 143K, Rows: 22M (NO STATISTICS)] (PATH ID: 7)
Projection: A.A_SD_LINES_HASH_superMaterialize: a1.A_SD_LINES_ID, a1.actual_date, a1.HASH
+-- Inner -> JOIN HASH [LeftOuter] [Cost: 61K, Rows: 22M] (PATH ID: 8)
Join Cond: (a2.A_Date_ID = t2.sd_POSTINGDATE_A_Date_ID)
+-- Outer -> STORAGE ACCESS for t2 [Cost: 52K, Rows: 22M] (PATH ID: 9)
Projection: A.sd_POSTINGDATE_superMaterialize: t2.sd_POSTINGDATE_A_Date_ID
+-- Inner -> STORAGE ACCESS for a2 [Cost: 612, Rows: 4K] (PATH ID: 10)
[20032] is a SSIS error.
Is the query running from SSIS showing up in the DC_CANCELS data collector table when it fails?
See: https://www.vertica.com/blog/display-canceled-queries-quick-tip/
Maybe a network timeout is killing the query?
See: https://www.vertica.com/blog/configuring-tcp-idle-settings-long-running-idle-sessions/