quick overview of Email Server configuration:
Message Labs 3rd party spam filter scans and sends email through
VMware Virtualized on a blade server cluster connected to dedicated iSCSI storage.
All vmware drives are thick provisioned
2 CAS Servers Load balanced ~ 100gb C: (40% full)
2 MBX Servers with a DAG. 100gb C: (50% full) 2TB D: (10% full) for mailbox databases and 500gb E: for transaction logs. Logs get cleared after successful exchange aware backup.
My Question is this.
Backups started but ended up filling the C: on the mailbox database servers which obviously caused those MDB servers to stop working. Since the CAS servers were still up and running, would inbound email queue up at the CAS Servers being unable to deliver it at all to a database server (i know its supposed to act simply as a proxy) or would it try to deliver them knowing the recipient does exist and hence the email is lost? or would the end user get a bounce back, undeliverable? Ive got a decent understanding of how the mail flow inbound works but just curious about this specific instance. I had a 3 hour window, just under 1 during normal business hours.