Symptoms
MCP cores when importing a configuration with 12000 nested address lists and auto-sync enabled.
/var/log/ltm contains messages similar to the following:
-- err clusterd[7274]: 013a0004:3: IO error on recv from mcpd - connection lost
-- info sod[7953]: 010c0009:6: Lost connection to mcpd - reestablishing.
-- notice chmand[7594]: 012a0005:5: resetting chmand services
-- err snmpd[7952]: 010e0001:3: Cannot communicate with MCPD server.
-- err mysqlhad[7596]: 014e0006:3: MCP Failure: 1.
-- err zxfrd[7962]: 0153e0f7:3: Lost connection to mcpd.
-- err tmrouted[6299]: 01910013:3: FATAL error: 6 irrecoverable MCP I/O error (Unknown error 16908291).
--
err alertd[7280]: 01100042:3: Failed with MCPD at: MCP msg receive (16908291).
-- err alertd[7280]: 01100042:3: Failed with MCPD at: Socket read (16908291).
Conditions
-- AFM configuration.
-- Devices in a device group trust configuration.
-- Device group configured with Autosync enabled.
-- Importing a configuration with a very large number of nested address lists (for example, 12000 nested address lists).
Workaround
Split the configuration into smaller chunks (e.g., 1000 address lists each) and load them one at a time.
