There are several factors which can contribute to these types of errors/symptoms, some include the following:
- Improper registry keys
- Out of date or poorly written NIC drivers
- NIC Teaming
- NIC Bindings
- OS patches or service packs which have made changes to TCP/IP settings
- TCP Chimney Offload and/or TCP/IP Offload Engine (TOE) are enabled (See Related Documents at bottom of this article)
- I/O degradation on indexing storage device
- Size of Index
- Schema Type of Index
(Symantec Article: TECH62307) for troubleshooting information.
From a support forum: "We have implemented all of the index best practices, have excluded all the pertinent directories from virus scanning, and do not have the 3gb switch set on any of the servers.
We found the index corruption occurred weekly right around the time our servers came up after a reboot. So we believe the problem is from not waiting long enough after we stop the EV services to reboot the server. Symantec support told us that anytime we reboot the server we should stop all EV services, wait 5-10 minutes, then reboot.
The reason for this is if you just reboot the server then the indexing service may not have time to finish committing what it has in memory. If the indexing service cannot finish its process then it cannot make the index stable and causes index corruption.
It seems like this is more of an issue with the newer versions of EV because we never had this problem before."