When performing mass update operations execution order of update SQL statements is not ordered by itemid in the frontend.

That is why can appear many failed queries:

"Query failed: [0] PGRES_FATAL_ERROR:ERROR: current transaction is aborted, commands ignored until end of transaction block"

Seems that this appears due to the fact that execution order of these update sql statements is not ordered by itemid:

update items set lastclock=1304516778,prevvalue=lastvalue,prevorgvalue=NULL,lastvalue='1' where itemid=83278;
...