We've been battling with this problem for months and months - it's very intermittant and would affect random mailboxes.
The only way we found around it was exporting the account via exmerge (and taking all details re DLs, delegates, rules, etc), deleting and recreating the mailbox and then using exmerge to import the data into the new mailbox (and resetting everything up, synchronising, etc). This worked fine but was really time consuming.
About a month ago MS found a quicker fix for us. We have been testing this over the last couple of weeks and have found it to be really good.
The tool we have been given and have been testing is called "DamagedRuleScan". It can be run from a Workstation. We are an NT 4.0 site running Exchange 5.5 but this tool can only run from a Win2K sp2 pc and I believe can also be run against Exchange 2000. It is apparently an internal MS tool written by one of their engineers and comes with the same disclaimer as a resource kit tool - eg not supported, etc. Anyway it works for us - pretty happy.
Alternatively we were advised as follows:
"If however you prefer the manual method you could always turn up diagnostic logging on the Rules objects on the Private Information Store you will find that when OOF/rules are broken (in an error state) the following event will fire:
Event id: 1035
Source: MSExchangeIS Private
Category: Rules
Description: The Rule 1-7E2 with sequence number 50 is being ignored because it is in an error state. The distinguished name of the owning mailbox is [mailbox name]
This event will be generated when a rule is suppose to fire but doesn’t, hence you will be in a reactive mode. From this point it’s the manual process to determine if the rule is an OOF, then delete the OOF message and recreating it again [using something like Mdbvu32.exe]."
Hope this helps some
Kay
