Quantum| Tape Alert Overview |
The HP Tape Alert Specification from the t10.org SCSI committee Web site provides a good overview, including the usage, format, and function of Tape Alerts. To download this document, use this link: http://www.t10.org/ftp/t10/document.02/02-142r0.pdf.
The discussion below provides more information about when Tape Alerts occur, why they occur, and what they mean, beyond the overview given in the document listed above.
A Tape Alert is an event composed of data that a supporting device presents upon request from an application via a SCSI Log Sense Select page. Tape Alert data can be presented/supported by a number of different devices. In most Quantum deployments, Tape Alert data is accessed through a Tape Library or through Tape Drive devices. The Tape Alert Specification outlines ways in which an application can access/request Tape Alert data from a device at certain times, after confirming that the device supports Tape Alert (using Log Sense page 0x2e).
The specification states that at a minimum, Tape Alert data can be read from a "Tape Drive" at the following times:
The specification states that Tape Alert data can be read from a "Library" at the following times:
Some backup applications periodically "poll" devices for Tape Alert data. The specification allows polling for Tape Alert data to occur every 60 seconds, but this is not a requirement. Though not mandatory, the host software may also poll the Log Sense page every 60 seconds while the tape drive/library is idle.
There are a total of 64 Tape Alert "flag" definitions that are checked. Each Tape Alert flag has its own unique definition, causes, severities, and recommended actions. These are discussed in a separate document in the "Tape Alert Definitions" section of this article.
Events that induce Tape Alerts often trigger more than one at a time. For example, when a drive/media event involving a read/write error that triggers a "Hard Error" Tape Alert 3 occurs, it may also trigger a "Read Error" Tape Alert 5, and a "Clean Now" request via a Tape Alert 20. It is important to understand the multiple Tape Alerts tied to a single event, so that accurate trends can be defined when deriving corrective actions. Once a trend of a specific Tape Alert is established, the corrective actions recommended for the respective Tape Alert can be applied.
The Tape Alert data is reset when:
There are three categories of Tape Alert flags:
Note that depending on the severity of the Tape Alert, a backup application or device that logs the Tape Alert data may archive that data differently, based on the severity. For example, on Quantum Scalar i2000 firmwares before i6.5, certain Tape Alerts categorized as informative were logged into the library's logging archive (the rmTapealert.cvs file). However, in Scalar i2000 FW above i6.5, certain Tape Alerts with the severity set to informative are no longer logged. This may explain some differences that might be seen when comparing a backup application's Tape Alert logging history with the Tape Alert logging history from a library.
To understand what each tape alert means, and what actions to perform to mitigate the issue, see the complete listing of tape alert definitions from CSWeb.
| Notes |
This page was generated by the BrainKeeper Enterprise Wiki, © 2018 |