The execution time of agent 'Vamsoft ORF SMTP Receive Agent' exceeded 90000 milliseconds RSS Back to forum
@christopher.low:
The error message indicates that the Transport Agent of ORF could not finish the tests at the On Arrival (OnEndOfData) filtering point within 1.5 minutes. We have seen such long processing times caused by the tarpit delay (http://vamsoft.com/r?o-hto-adm-tarpitdelay-settings) - it has the exact purpose of making processing long.
Assuming you have tarpit delay enabled, I recommend checking your related settings (Filtering / Actions, Edit button under On Arrival, Tarpit delay at the bottom): it is recommended to be configured to "Delay response on recipient validation or Recipient Blacklist hit only" and the delay should be a minute or less. If setting these values does not solve the problem, I recommend disabling Tarpit Delay entirely and saving your settings by pressing Ctrl + S to apply the changes.
When such timeout occurs, ORF may not log the email and but let Exchange deliver it, so you will not lose any emails because of this (hence the "However, Transport will continue processing this message" message). However, the email flow will be slowed down so much it appears it has been stopped - each tarpitted email will keep an SMTP connection open, and your server eventually runs out of threads, so newly arrived emails will have to wait until the tarpit period expires and an SMTP connection is freed up.
the filter on arrival doesn't enable tarpit. I suppose it was a glitch.
Just had absolutely the same symtoms on newly installed Windows Server 2008 R2 Enterprise and Exchange 2010 Standard with ORF Fusion 5.1
Same error strated to appear in the event log, at the same time incoming mail stopped until the server was restarted (after ~1 hour).
Tarpit delay is not enabled.
What else could be the cause?
@alex.lutikov:
Anything really that causes ORF to respond slower than 90 seconds. The typical culprit is a non-default Tarpit Delay, but quite often the issue is with DNS. For instance, let's assume that your have 3 DNS servers configured in ORF and 60 seconds timeout. If the DNS zone being looked up by ORF is non-responsive (e.g. DNS server went down), it may very well take 60 * 3 = 180 seconds before ORF gives up using all three servers. More commonly, one of the DNS servers specified for ORF goes down and this may have similar effects.
A few other things that may cause similar issues are:
* Database timeouts when using External Databases: ORF employs no timeout controls on these, but you can configure it in the connection string if supported by the SQL client.
* External Agent timeouts: ORF does use timeout control here, but then these may be increased in the External Agent definition.
* Private Local Database corruption: even though timeout control is employed, a corrupted database can do nasty things like deadlocks, causing ORF to become partially unresponsive.
Just a fyi, I will try to see if the problem repeats
The execution time of agent 'Vamsoft ORF SMTP Receive Agent' exceeded 90000 milliseconds while handling event 'OnEndOfData' for message with InternetMessageId: '<001901cdee3b$bbbd7810$94b025dd@LocalHostpnx2s>'. This is an unusual amount of time for an agent to process a single event. However, Transport will continue processing this message.
--
just happened today. the agent hanged or something so that all incoming mail stopped for over an hour before users complained. resolved after rebooting the mail server, I did not try restarting the service though.