Regardless of the degree of complexity in the underlying SAN disk architecture, the really important measure is the service that the SAN provides to its ‘client’ servers. Metron’s Athene provides measurements of read and write response times to logical and physical disks.

Recent studies have highlighted occasions when, whilst SAN read performance was satisfactory, very long response times were recorded for writes during periods of high write activity. If you are getting similar problems – check the write cache batteries!

Most SANs provide a mirrored battery-backed write cache. The batteries are to ensure that disk write data is not lost in the case of a power failure. If the battery charge falls below certain thresholds (or, as was the case in one example, the batteries are missing!) then the SAN system will disable the write cache to ensure absolute data integrity. The effect of this can be dramatic, with response times changing from a few milliseconds, to tens or even hundreds of milliseconds per write.

Using Athene’s reporting capabilities, you can check for this very important metric, and raise the alarm if response times suddenly degrade

To Management Performance Tips Index