Notes.ini Entry
Name:
RouterDisableFaultDataCapture
Syntax
RouterDisableFaultDataCapture=
0 / 1
Applies to:
Servers
Add-on:
First Release:
9.0.1
Obsolete since:
Category:
Crash, Router
Default:
None
UI equivalent:
None
Description:
Setting this Variable to 1 will result in the creation of a pointer to a single message that caused a crash and at restart identifying and movement of the bad Message.
Setting this Variable to 0 or "" will result in the previous behaviour, and could make the server crash repeat itself over and over again.
Protect from repeat outages due to a single message instance:
Processing a “bad” message is responsible for the crash. It remains in mail.box or mail file. Server restarts and proceeds to deal with the same “bad” message, causing repeat crash for below scenarios:
- Transfer or deliver same "bad" message: Router repeat crash
- Receive same "bad" message: SMTP repeat crash
- Fetch same “bad” message via IMAP: IMAP repeat crash
Solution to give messaging server reliability in 9.0.1:
- Keeping per-thread context identifying the current message being processed.
- Registering an exception handler callback that is called at time of crash to record which email message was processed during crash. A data file is opened and the information identifying this message is written to the file.
- When server restarts, if above file exists, Router/SMTP/IMAP read the file to identify the "bad" message, move the "bad" message to a new DB (Router and IMAP), and continue to deal with remaining messages.
Router diagnostic data provides additional information in NSD stacks
- This identifies work in progress by a router transfer/delivery thread at the time of a crash.
The information includes
- Message being processed (mailbox and Note ID)
- Sender and recipient.
- Stacks in the NSD contain a string printing out this value.