Server Crashing

The MailArchiva process exits unexpectedly or the web console is unresponsive.

Incorrect amount of heap space

The amount of heap space is inappropriate for your hardware configuration / workload.  MailArchiva requires enough heap space and virtual memory available. The amount of heap space is set by the person installing the software as part of the install process. The set value can be changed at any time. However, a restart of the MailArchiva service is required after changing heap space settings.

Assuming there is 8GB of physical memory available:

Example 1) Assign 7GB of heap space: crash probable due to only 1GB virtual memory available for the server (memory outside the heap) to allocate

Example 2) Assign 512MB of heap space: crash probable, esp. when the server is subject to heavy workload or encounters large attachments for indexing, due to insufficient heap space available

Example 3) Assign 4GB of heap space: this setting is good for 200 mailboxes, as there is plenty of heap space and virtual memory (memory outside of the heap) available.

Assuming there is 128GB of physical memory available:

Example 1) Assign 64GB of memory: CPU resources are required to manage a large amount of memory. If the server only had 2 processing cores, the server could become unresponsive. When allocating a large amount of heap (e.g. >6GB), it is advised to have a beefed up machine with plenty of CPU resources.

To change the heap space. On Linux type /opt/mailarchiva/server/configure. On Windows, open the MailArchiva Service Manager from Start Menu, right click Service Manager icon from Task Tray, click Configure, click Java tab, set Maximum Memory Pool value to the appropriate value. Refer to Memory Settings for more info. After changing the heap space settings, a restart is required.
 

Corrupted receive / index queue

If the server has a corrupted receive or index queue, the server may become unresponsive. Take the following steps to resolve this issue:

1) Stop MailArchiva service
2) Rename /var/opt/mailarchiva/ROOT/queue/receive to  /var/opt/mailarchiva/ROOT/queue/restore_receive (Linux)

    Rename C:\ProgramData\MailArchiva\ROOT\queue\receive to C:\ProgramData\MailArchiva\ROOT\queue\restore_receive (Windows)

3) Rename /var/opt/mailarchiva/ROOT/queue/index to  /var/opt/mailarchiva/ROOT/queue/restore_index (Linux)

     Rename C:\ProgramData\MailArchiva\ROOT\queue\index to C:\ProgramData\MailArchiva\ROOT\queue\restore_index (Windows)

4) Start MailArchiva service

Corrupted graph database

If there is a corrupted Graph DB (Orient DB), the server may stop archiving or behave in a problematic manner. Refer to fix graph DB for a resolution.

Insufficient / failing hardware resources

Not enough disk space - If the server runs out of disk space, the server may become unresponsive or the behaviour of the server may become erratic. Please ensure there is enough disk space on all partitions. 

Not enough CPU resources - There may not be enough CPU resources to handle the workload assigned to MailArchiva. The resolution is to add more CPU core to the virtual machine, install a new CPU or change processors.

Failing network attached storage - If the server is archiving to network attached storage (NAS), and the NAS becomes unresponsive or there are frequent network I/O errors, the server can freeze up. Of particular note, it is not advised to place MailArchiva volume indexes on a remote storage accessible via SMBFS. Indexing requires low latency drives.

Faulty memory, CPU processors or motherboard - One cannot rule out the possibility that there is a problem with the hardware itself.