[Bug #207482, 238785]
Pages on disk may be corrupted if:
The write I/O requested by Adaptive Server is 8K or more, either because of using larger buffer pools or because of large database page size, and
You are using asynchronous I/O in a UNIX file system (UFS) on Solaris, and
Adaptive Server crashes or is shut down using the shutdown with nowait option under heavy I/O conditions.
This error is caused by a Sun operating system problem. The Sun service order number is 4226801. When Sun corrects the problem, the workarounds will be unnecessary.
Workaround:
Do not use asynchronous I/O with UFS on Solaris. Instead, use asynchronous I/O with raw partitions, or
Use Adaptive Server 2K database page size and named caches only with less than 8K buffer pools.