xref: /illumos-gate/usr/src/cmd/fm/eversholt/files/common/disk.esc (revision bbf21555)

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright 2016 Nexenta Systems, Inc.  All rights reserved.
 */

#pragma dictionary "DISK"

#define	P			disk

fru P;
asru P;

/*
 * Over all comments for this file:
 * <disk-as-detector> The disk-as-detector DE provides the mapping between
 * ereports generated by a kernel disk driver sd(4D) and resulting faults.
 */

/*
 * SERD engine for media error fault propagation:
 *
 * This strategy is designed to give a file system, like ZFS, the
 * ability to attempt data recovery/relocation without faulting a disk.
 * This implementation depends on a file system retry to the same lba
 * to trigger a fault when recovery/relocation is not possible.
 *
 * We let the engine propagate one error only once every 1 minute and then if we
 * still get 2 or more errors within 24 hours for the same LBA,
 * there is a fault.
 */
engine serd.io.scsi.cmd.disk.dev.rqs.merr@P, N=1, T=24h;

/*
 * disk-as-detector: fault events.
 */
event fault.io.scsi.cmd.disk.dev.rqs.derr@P;
event fault.io.scsi.cmd.disk.dev.rqs.merr@P,
    engine=serd.io.scsi.cmd.disk.dev.rqs.merr@P;

/*
 * The uderr fault will be defined at some future time.
 * event fault.io.scsi.cmd.disk.dev.uderr@P;
 */

/*
 * disk-as-detector: upset events.
 * NOTE: For now we define an upset to implement discard.
 */
event upset.io.scsi.cmd.disk.dev.rqs.derr@P;
event upset.io.scsi.cmd.disk.dev.rqs.merr@P;
event upset.io.scsi.cmd.disk.dev.uderr@P;
event upset.io.scsi.cmd.disk.dev.serr@P;
event upset.io.scsi.cmd.disk.tran@P;
event upset.io.scsi.cmd.disk.recovered@P;

/*
 * disk-as-detector: ereports from the kernel.
 *
 * We don't know the topology for all scsi disks, but the kernel will always
 * generate ereport telemetry assuming that we do. We define these ereports
 * with 'discard_if_config_unknown=1', which permits ereports against things
 * with unknown topology to be silently discarded.  The ereport data is logged
 * in either case, and can be viewed via 'fmdump -eV'.
 */
event ereport.io.scsi.cmd.disk.dev.rqs.derr@P, discard_if_config_unknown=1;
event ereport.io.scsi.cmd.disk.dev.rqs.merr@P, discard_if_config_unknown=1;
event ereport.io.scsi.cmd.disk.dev.serr@P, discard_if_config_unknown=1;
event ereport.io.scsi.cmd.disk.dev.uderr@P, discard_if_config_unknown=1;
event ereport.io.scsi.cmd.disk.recovered@P, discard_if_config_unknown=1;
event ereport.io.scsi.cmd.disk.tran@P, discard_if_config_unknown=1;

/*
 * For some ereports we let the 'driver-assessment', communicated as part of
 * the ereport payload, determine fault .vs. upset via propagation constraints.
 */
#define DRIVER_ASSESSMENT_FATAL		\
	    (payloadprop_contains("driver-assessment", "fatal"))
#define DRIVER_ASSESSMENT_NONFATAL	(!DRIVER_ASSESSMENT_FATAL)

/*
 * disk-as-detector: propagations from faults(based on
 * DRIVER_ASSESSMENT_FATAL).
 * We need to set additional fault payloads to indicate fault details.
 * The payload we may need are listed as following:
 * fault.io.scsi.cmd.disk.dev.rqs.derr
 *     op_code, key, asc, ascq
 * fault.io.scsi.cmd.disk.dev.rqs.merr
 *     op_code, key, asc, ascq, lba
 */
prop fault.io.scsi.cmd.disk.dev.rqs.derr@P->
    ereport.io.scsi.cmd.disk.dev.rqs.derr@P{ DRIVER_ASSESSMENT_FATAL &&
    setpayloadprop("key", payloadprop("key")) &&
    setpayloadprop("asc", payloadprop("asc")) &&
    setpayloadprop("ascq", payloadprop("ascq"))};

/*
 * Utilize setserdsuffix with specific LBA,
 * the serd engine would only trigger if the fault recurred on the same LBA
 */
prop fault.io.scsi.cmd.disk.dev.rqs.merr@P->
    ereport.io.scsi.cmd.disk.dev.rqs.merr@P{ DRIVER_ASSESSMENT_FATAL &&
    setserdsuffix(payloadprop("lba")) &&
    setpayloadprop("key", payloadprop("key")) &&
    setpayloadprop("asc", payloadprop("asc")) &&
    setpayloadprop("ascq", payloadprop("ascq")) &&
    setpayloadprop("lba", payloadprop("lba"))};

/*
 * NOTE: this propagation uses the "may" propagation of eversholt.
 * The ereport need never exist. It's just a way of making
 * the diagnosis wait for the within time on that ereport
 * to complete. Once it has completed the diagnosis continues
 * even though the dummy ereport didn't occur.
 */
event ereport.io.scsi.cmd.disk.dev.rqs.merr.dummy@P {within(60s)};
prop fault.io.scsi.cmd.disk.dev.rqs.merr@P (0) ->
	ereport.io.scsi.cmd.disk.dev.rqs.merr.dummy@P;

/*
 * The uderr fault will be propagated at some future time.
 * prop fault.io.scsi.cmd.disk.dev.uderr@P->
 *     ereport.io.scsi.cmd.disk.dev.uderr@P{ DRIVER_ASSESSMENT_FATAL };
 */

/*
 * disk-as-detector: propagations from upsets(based on
 * DRIVER_ASSESSMENT_NONFATAL).
 */
prop upset.io.scsi.cmd.disk.dev.rqs.derr@P->
    ereport.io.scsi.cmd.disk.dev.rqs.derr@P{ DRIVER_ASSESSMENT_NONFATAL };

prop upset.io.scsi.cmd.disk.dev.rqs.merr@P->
    ereport.io.scsi.cmd.disk.dev.rqs.merr@P{ DRIVER_ASSESSMENT_NONFATAL };

/*
 * disk-as-detector: propagations from upsets(independent of
 * driver-assessment)
 */

prop upset.io.scsi.cmd.disk.dev.serr@P->
    ereport.io.scsi.cmd.disk.dev.serr@P;

prop upset.io.scsi.cmd.disk.dev.uderr@P->
    ereport.io.scsi.cmd.disk.dev.uderr@P;

prop upset.io.scsi.cmd.disk.recovered@P->
    ereport.io.scsi.cmd.disk.recovered@P;

prop upset.io.scsi.cmd.disk.tran@P->
    ereport.io.scsi.cmd.disk.tran@P;

/*
 * --------------------------------------
 * The remainder of this file contains rules associated with the operation of
 * cmd/fm/modules/common/disk-monitor/disk_monitor.c code.
 *
 * The disk DE provides a very simple 1-to-1 mapping between SCSI disk events
 * generated by the disk-transport fmd module, and the resulting faults.
 */

/*
 * Fault events.
 */
event fault.io.disk.over-temperature@P,
    FITrate=10, FRU=P, ASRU=P;
event fault.io.disk.predictive-failure@P, FITrate=10,
    FITrate=10, FRU=P, ASRU=P;
event fault.io.disk.self-test-failure@P, FITrate=10,
    FITrate=10, FRU=P, ASRU=P;
event fault.io.disk.ssm-wearout@P;

/*
 * ereports.
 */
event ereport.io.scsi.disk.over-temperature@P;
event ereport.io.scsi.disk.predictive-failure@P;
event ereport.io.scsi.disk.self-test-failure@P;
event ereport.io.scsi.disk.ssm-wearout@P;

/*
 * Propagations.
 */
prop fault.io.disk.over-temperature@P ->
    ereport.io.scsi.disk.over-temperature@P;

prop fault.io.disk.self-test-failure@P ->
    ereport.io.scsi.disk.self-test-failure@P;

prop fault.io.disk.predictive-failure@P ->
    ereport.io.scsi.disk.predictive-failure@P {
    setpayloadprop("asc", payloadprop("additional-sense-code")) &&
    setpayloadprop("ascq", payloadprop("additional-sense-code-qualifier")) };

prop fault.io.disk.ssm-wearout@P ->
    ereport.io.scsi.disk.ssm-wearout@P {
    setpayloadprop("current-wearout-percentage",
    payloadprop("current-ssm-wearout"))
    && setpayloadprop("threshold-wearout-percentage",
    payloadprop("threshold-ssm-wearout")) };