RU Contact us
MSX / 03 of 09 · Microsoft and Exchange

Mail that survives a server failure and an admin mistake

Exchange and Active Directory deployment, migration and recovery. We also take on the systems everyone is afraid to touch.

When people call us
  • 01The Exchange upgrade has been postponed for three years
  • 02A database stopped mounting after the disk filled up
  • 03Mail is delayed and queues grow for no visible reason
  • 04The migration stalled halfway through the mailboxes
+Microsoft and Exchange: ink illustration
DAG / witness on the same host as a node
LOG / transaction logs fill 94% of the disk
MSX / 03Findings are illustrative
+
Day one

What we check first

  • Check DAG health, database copies and quorum
  • Review transaction log growth and truncation after backup
  • Check DNS, certificates and Autodiscover
  • Assess Active Directory: replication, FSMO roles, sites
Tool

DAG quorum: how many failures it survives

Counts node and witness votes by the majority rule.

+
Failure patterns

Where it usually breaks

A

A quorum that cannot survive a failure

An even number of nodes with no witness, or a witness next to a node. The first failure stops the whole cluster.

B

Logs ate the disk

Backups do not complete, logs are not truncated, the disk fills up and databases dismount.

C

A migration with no way back

The hybrid setup was dismantled too early, and mailboxes can no longer be moved back.

+
Outcome

What you keep

  • A resilient design with a tested failover
  • A migration plan with checkpoints and rollback
  • Database and mailbox recovery procedures, tested in practice
← SEC

Information security

VRT →

Virtualization

+
Contact

Describe your task

Tell us what is happening and what has been tried. The engineer who will handle it replies.