sudden massive "Invalid Range from HWM for lbid" was thrown

At a certain point many
"CAL0000: Invalid Range from HWM for lbid .... range size should be <= blocksReadAhead"
will b thrown.

Issue is not reproducabe yet.

-For impacted tables a cpimport runs sometime before error popped up.

• error will thrown shortly, after a "select" will be executed on the impacted table.
• lbid pointed to a dictionary object_id of the impacted column
• the lbid in error message is not mentioned in any cpimport logs.
• primproc memory consumption is high

No crashes in folder traces, no suspicious errors are visible.

Issue happens on multinode cluster and singlenodes server.