Overview

When Exinda appliance is experiencing intermittent disconnections, the following log message can be generated:

[wmd.ERR]: std::exception::what: Failed to create database tables: disk I/O error {sql_exec@persist/PersistentStateDB.cpp:366}
[pm.NOTICE]: Process wmd (pid 14769) terminated from signal 6 (SIGABRT)

As a result, the wmd process is crashing and no longer functioning.

Placing Exinda into bypass mode resolves disconnection issues.

Solution

Because of the WAN Memory process crash, there could be issues accelerating traffic with the other site Exinda. Checking Configuration > System > Setup > Storage (or using CLI df -h command) shows the wmd partition is full and is causing the wmd process to crash:

The memory folder can be cleared by resetting WAN Memory:

1. Stop the Wan Memory service under Configuration > System > Optimization > Services
   
2. Clear (format) the Wan Memory partition under Configuration > System > Setup > Storage
   
3. Clear the Wan Memory cache under Configuration > System > Optimization > Wan Memory by clicking the Expire and Reset buttons
   
4. Restart the Wan Memory service under Configuration > System > Optimization > Services

Important: It's recommended to increase the size of the partition.