Category Archives: Hackery

One-liner: Check the five, most important SMART parameters on a disk.

A while ago, Backblaze published a report on what they consider to be the most reliable SMART parameters for determining whether a disk is failing. These include:

  • 5 – Reallocated_Sector_Ct
  • 187 – Uncorrectable_Error_Cnt
  • 188 – Command_Timeout
  • 197 – Current_Pending_Sector_Count
  • 198 – Offline_Uncorrectable

For a complete description of these parameters, take a look at the Wikipedia article on SMART.

While our sample of failing disks is no where near as large as Backblaze’s, our results have, unsurprisingly, correlated pretty strongly to theirs.

Note that not all of these parameters are supported by the drive manufacturers and that we typically don’t see many of these parameters on the hard disks supplied in Apple hardware. Additionally, note that SMART is not supported on some drives.

Assuming you’ve got smartmontools installed, this one-liner will very quickly give you a snapshot of the key values we look for as strong indicators that a drive needs to be replaced:

smartctl -a disk0 | egrep "^( 5|187|188|197|198)"

where

disk0

is the disk you’re testing. To get the disks available to test, run

diskutil list

You’ll get back output that looks like this:

/dev/disk0
   #:                       TYPE NAME                    SIZE       IDENTIFIER
   0:      GUID_partition_scheme                        *256.1 GB   disk0
   1:                        EFI                         209.7 MB   disk0s1
   2:                  Apple_HFS Macintosh HD                 255.2 GB   disk0s2
   3:                 Apple_Boot Recovery HD             650.0 MB   disk0s3
/dev/disk1
   #:                       TYPE NAME                    SIZE       IDENTIFIER
   0:      GUID_partition_scheme                        *500.1 GB   disk1
   1:                        EFI                         209.7 MB   disk1s1
   2:                  Apple_HFS Storage                 499.8 GB   disk1s2

In the example above, there are two disks to choose from,

disk0

and

disk1

Assuming the drive supports all five SMART parameters, you’ll get back something that looks like this:

  5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  Always       -       0
187 Reported_Uncorrect      0x0032   100   100   001    Old_age   Always       -       0
188 Command_Timeout         0x0032   100   100   001    Old_age   Always       -       0
197 Current_Pending_Sector  0x0032   100   100   001    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0030   100   100   001    Old_age   Offline      -       0

Those trailing zeros are what we like to see. Positive values in the last column mean that the drive probably needs to be replaced.

Sometimes, you stumble on the right person…

…and they reveal to you a bit of magic you didn’t know existed.

In this case, it is an undocumented flag in Promise Technology’s command line utility for the Promise Pegasus2 Thunderbolt RAID, promiseutil.

As previously discussed, it appeared to be impossible to check the status of more than one Promise Pegasus enclosure from inside a script using promiseutil. We had filed a support ticket, hoping for some kind of resolution, but were told that promiseutil works as intended.

On a hunch, I reached out to someone at Promise and asked for their help investigating this issue.  I was pleasantly surprised when the contact not only took the issue seriously, he immediately looped in other support engineers to look at the problem.

After a week of back and forth about what an appropriate solution would be, perhaps a feature request, the support engineer discovered that there is an undocumented flag that allows you to specify the hba of the Promise unit you want to execute a command on.

Here’s an example. Let’s say we want to check the SMART status of two Promise Pegasus from the command line:

 promiseutil -C smart -v

will return the information for the default device only.

If you want to be explicit about which Promise Pegasus you’re checking, first get the hba numbers of the connected units:

promiseutil -C spath

The results will be something like this:

archer:~ admin$ promiseutil -C spath
=================================================
Type  #    Model        Alias   WWN          Seq
=================================================
hba   1  * Pegasus2 R4       2000-0001-5558-2fe2  1
hba   2    Pegasus2 M4       2000-0001-5558-3f92  1

Now we use the magic (apparently undocumented) -P (uppercase, not the documented lowercase) flag to specify the unit we want to look at.

promiseutil -T hba -P 1 -C smart -v

which returns the results for the first unit.

promiseutil -T hba -P 2 -C smart -v

will return results for the second unit.

My sincere thanks to the people at Promise who helped us sort this out (you know who you are) and to my fellow bug wrestler, Allen Hancock of Watchman Monitoring.

As always, be cautious with promiseutil. Its power is mighty and Bad Things® can happen if used incorrectly.

Logging time-stamped ping results to a file using Applescript and bash.

I deal with a number of remote workers who, for one reason or the other, don’t work in the company office. Often, they’re using a VPN tunnel to connect to a server back at the company.

Occasionally, we’ll see intermittent connectivity issues from the client. Perhaps it’s their ISP, perhaps it’s the VPN tunnel, perhaps it’s a piece of software triggering IDS on a managed firewall.

In any case, we can triangulate the problem by launching a script on the client’s side that pings endpoints of our choosing to check connectivity. But we also want to time stamp and capture the results of the pings to a text file we can review later.

This is where

tee

is your friend. As the man entry says, tee is a “pipe fitting”.

The tee utility copies standard input to standard output, making a copy in zero or more files.

So, here are our requirements:

  1. Script is user-initiated.
  2. Script gets out of the user’s way.
  3. Timestamps and logs the pings to a text file in a  folder on the Desktop.

This Applescript, which makes a bunch of bash calls, does all of that.

# Simple ping monitor
# A script that pings servers of your choice by IP or DNS name and logs the results to a text file in a folder on the Desktop.
#
# Written by AB @ Modest Industries (modestindustries.com)
#
# 2012-07-25 - AB: First draft.
# 2014-07-25 - AB: Formatting cleanup. 

#Servers to ping. For each server you name here, you'll need to set up a ping statement below.
set server1 to "google.com"
set server2 to "8.8.8.8"
set server3 to "yahoo.com

property the_prefix : space

property the_sep : "-"

# Format a date to use as a datestamp.
on myDate()
    
    set myYear to "" & year of (current date)
    
    set myMth to text -2 thru -1 of ("0" & (month of (current date)) * 1)
    
    set myDay to text -2 thru -1 of ("0" & day of (current date))
    
    set myHours to hours of (current date)
    
    set myMinutes to minutes of (current date)
    
    return {myYear, myMth, myDay, myHours, myMinutes}
    
end myDate

# Check for a folder called Monitoring on the Desktop. If it doesn't exist, make one.
tell application "Finder"
    set the directory to desktop
    if (exists folder "Monitoring") is false then
        make new folder at desktop with properties {name:"Monitoring"}
    end if
    
    set the_path to folder "Monitoring" of desktop
    
    set the_name to (item 1 of my myDate())
    
    set the_name to (the_name & the_sep & item 2 of my myDate())
    
    set the_name to (the_name & the_sep & item 3 of my myDate())
    
    set the_timestamp to item 4 of my myDate() & item 5 of my myDate()
    
    -- set the directory to "Monitoring"
    if (exists folder the_name of folder "Monitoring" of desktop) is false then
        make new folder at the_path with properties {name:the_name}
    end if
    
    set the_path to folder the_name of folder "Monitoring" of desktop as alias
    
    set posixPath to POSIX path of the_path
end tell

# Ping servers of your choice. You'll need one statement for each server named above.

tell application "Terminal" to do script "ping " & server1 & " | while read pong; do echo \"$(date): $pong\"; done | tee " & quoted form of posixPath & the_name & the_sep & the_timestamp & the_sep & server1 & ".txt"

tell application "Terminal" to do script "ping " & server2 & " | while read pong; do echo \"$(date): $pong\"; done | tee " & quoted form of posixPath & the_name & the_sep & the_timestamp & the_sep & server2 & ".txt"

tell application "Terminal" to do script "ping " & server3 & " | while read pong; do echo \"$(date): $pong\"; done | tee " & quoted form of posixPath & the_name & the_sep & the_timestamp & the_sep & server3 & ".txt"

# Hide all the windows.
tell application "System Events" to set visible of process "Terminal" to false

# Tell the user it's running.
display dialog "Ping monitor is running!" buttons {"OK"} default button 1

# Switch back to the Finder.
tell application "Finder" to activate

You might want to tweak the dialogue to tell the user to leave the Terminal app running.

Should this be a bash script? Probably. But this works and can be launched by the user and hides most of the gubbins so that the user can get on with their business.

Promise Pegasus2: Scripting a SMART check with promiseutil

We’ve found that the Promise Pegasus2 Thunderbolt 2 RAID can report that the SMART Health status of its disks is just dandy, while the unit is quietly accumulating ATA errors that may indicate the pending failure of a disk.

I want to be notified if the Pegasus either has a SMART status failure or if ATA errors are present on any of the disks.

This script does just that. It’s essentially a more refined version of the previous promiseutil scripts that grabs the simple SMART status of each disk, greps to see if it’s “OK”, then runs a line of awk that looks at the report to see if there’s an “ATA Error Count”. As always, it logs to system.log and optionally sends error reports by email.

#!/bin/bash
#
# promise_smart_check.sh
#
# Checks Promise Pegasus2 SMART status, checks for ATA errors, logs and mails the output if there's an issue.
#
# Author: AB @ Modest Industries
#
# Requires Promise Utility for Pegasus2 (http://www.promise.com), tested with v3.18.0000.18
# Requires sendemail for email alerts (http://caspian.dotconf.net/menu/Software/SendEmail/)
#
# Edit History
# 2014-04-21 - AB: Version 1.0.
# 2014-04-24 - AB: Refactored.
# 2014-05-01 - AB: Incorporate the awk script to check for ATI errors.
# 2014-05-08 - AB: Refinements.
# 2014-05-15 - AB: Update to message body construction, tmp file & sendemail sanity checks.
# 2014-05-17 - AB: Added promiseutil path check.

export DATESTAMP=`date +%Y-%m-%d\ %H:%M:%S`

# Editable variables

# Path to sendemail
sendemail_path="/Library/Scripts/Monitoring/sendemail"

# Send email alerts?
send_email_alert=true

# Variables for sendemail
# Sender's address
alert_sender="[email protected]"

# Recipient's addresses, comma separated.
#alert_recipient='[email protected], [email protected]'
alert_recipient="[email protected]"

# SMTP server to send the messages through
alert_smtp_server="smtp.example.com"

# ------------ You probably shouldn't edit below this line ------------------
# Variables

# Default the error flags to false.
smart_error_flag="false"
ata_error_flag="false"

# Alert subject
alert_subject="ALERT: Promise Pegasus2 SMART problem detected on $HOSTNAME."

# Alert header
alert_header="At $DATESTAMP, a problem was detected on this device:\n"

# Pass / Fail messages
pass_msg="Promise Pegasus SMART check successful."
fail_msg=" *** Promise Pegasus SMART check FAILED!!! ***"

# Default the message body
message_body=""

# Alert footer
alert_footer="Run 'promiseutil -C smart -v' for more information."

# Promise Pegasus command line utility default path
promiseutil_path="/usr/bin/promiseutil"

# ----------------- Check for promiseutil, sendemail & set up temp files ------------------
if [ ! -f $promiseutil_path ]; then
        echo "$0 ERROR: $promiseutil_path does not exist"
        echo  "Please download and install the Promise Pegasus Utility app from http://promise.com"
        exit 1
fi

if [ ! -f $sendemail_path ]; then
        echo "$0 ERROR: $sendemail_path does not exist"
        echo  "Please download from http://caspian.dotconf.net/menu/Software/SendEmail/ and then set the \$sendmemail_path variable inside this script"
        exit 1
fi

unit_ID_tmp=`mktemp -q "/tmp/$$_unit_ID.XXXX"`
if [ $? -ne 0 ]; then
        echo "$0: ERROR: Can't create temp file, exiting..."
        exit 1
fi

smart_results_tmp=`mktemp -q "/tmp/$$_smart_results.XXXX"`
if [ $? -ne 0 ]; then
        echo "$0: ERROR: Can't create temp file, exiting..."
        exit 1
fi

# ----------------- Run promiseutil, evaluate the results ------------------

# Get Unit ID information for this Promise unit. Includes workaround for promiseutil tty issue.
screen -D -m sh -c "$promiseutil_path -C subsys -v >$unit_ID_tmp"

# Drop the output into a variable.
unit_ID=$(<$tmpdir$unit_ID_tmp)

# Get the SMART report, put it into a tmp file.
screen -D -m sh -c "$promiseutil_path -C smart -v >$smart_results_tmp"

# Grab the header for each PdId in the Promise
smart_status=$(cat $smart_results_tmp | grep -A4 "^PdId")

# Check the header to see if SMART Health Check reports a problem
if grep "^SMART Health Status:" <<< "$smart_status" | grep -qv "OK"
then
        smart_error_flag="true"
fi

# Check for ATA errors, which may indicate that the drive is failing even if SMART Health is OK
ata_errors=$(awk '/^PdId: [1-9][0-9]*/ \
                                { a=$0; n=4; next } \
                                n { --n; a=a "\n" $0; next } \
                                /^ATA Error Count*/ \
                                { ata_err=$0; print a "\n" ata_err "\n" }' \
                                "$smart_results")
# Flag if there were ATA errors
if [ "$ata_errors" != "" ]; then
        ata_error_flag="true"
fi

# ----------------- Build the message_body ------------------

# If there's a problem, build the header.
if [ "$smart_error_flag" ==  "true" ] || [ "$ata_error_flag" == "true" ]; then
        message_body="$alert_header\n\n$fail_msg\n\n$unit_ID\n\n"

        # SMART Health status.
        if [ "$smart_error_flag" == "true" ]; then
                message_body="$message_body\nSMART Health Status is reporting one or more bad drives."
        fi

        # Always include the smart_status
        message_body="$message_body\n\n$smart_status"

        # Then the ATA errors.
        if [ "$ata_error_flag" == "true" ]; then
                message_body="$message_body\n\nOne or more drives has an ATA Error Count and may be failing.\n\n$ata_errors"
        fi
fi

#  ----------------- Logging & email ------------------

# Log the results, conditionally send email on failure.
if [ "$ata_error_flag" == "true" ] || [ "$smart_error_flag" == "true" ]; then
        message_body="$message_body\n\n$alert_footer"
        echo "$DATESTAMP: \n\n$message_body" >> /var/log/system.log
        if [ "$send_email_alert" == "true" ] ; then
                "$sendemail_path" -f $alert_sender -t $alert_recipient -u $alert_subject -m "$message_body" -s $alert_smtp_server
        fi
else
        echo "$DATESTAMP: $pass_msg\n\n$unit_ID" >> /var/log/system.log
fi

# ----------------- Cleanup ------------------

rm -f rm -f $unit_ID_tmp $smart_results_tmp


This version of the script checks for the presence of promiseutil and sendemail. We call screen here because the promiseutil seems to need a TTY in order to run properly.

Hope you find it useful.

Promise Pegasus2: Scripting an Enclosure check with promise_enclosure_check.sh

The Promise Pegasus2 has onboard sensors that monitor the power supply  voltages, speed of the fan, and temperature of the controller and backplane.

This seems worth performing the occasional check on.

The example script below runs an initial check of the enclosure using promiseutil. If it doesn’t find that “Everything is OK”, it runs a more verbose check, logs the problem and optionally sends email.

#!/bin/bash
#
# promise_enclosure_check.sh
#
# Checks the status of a Promise Pegasus2 RAID enclosure and mails the output if there's an issue.
#
# Author: AB @ Modest Industries
#
# Works with Promise Utility for Pegasus2 v3.18.0000.18 (http://www.promise.com)
# Requires sendemail for email alerts (http://caspian.dotconf.net/menu/Software/SendEmail/)
#
# Edit History
# 2014-04-21 - AB: Version 1.0.
# 2014-05-08 - AB: Refinements.
# 2014-05-09 - AB: Better message_body if failed.

export DATESTAMP=`date +%Y-%m-%d\ %H:%M:%S`

# Editable variables

# Path to sendemail
sendemail_path="/Library/Scripts/Monitoring/sendemail"

# If a problem is found, send email?
send_email_alert=true

# Variables for sendemail
# Sender's address
alert_sender="[email protected]"

# Recipient's addresses, comma separated.
#alert_recipient='[email protected], [email protected]'
alert_recipient="[email protected]"

# SMTP server to send the messages through
alert_smtp_server="smtp.example.com"

# ------------ Do not edit below this line ------------------
# Variables

# Pass / fail flags
enclosure_pass=true

# The subject line of the alert.
alert_subject="Alert: Promise Pegasus2 enclosure problem detected on $HOSTNAME."

# Alert header
alert_header="At $DATESTAMP, an enclosure problem was detected on this device:\n"

# Pass / Fail messages
pass_msg="Promise Pegasus Enclosure check successful."
fail_msg=" *** Promise Pegasus Enclosure check FAILED!!! ***\n\n"

# Alert footer
alert_footer="Run 'promiseutil -C enclosure -v' for more information."

# Create temp files
unit_ID_tmp=`mktemp "/tmp/$$_unit_ID.XXXX"`
enclosure_results_tmp=`mktemp "/tmp/$$_enclosure_results.XXXX"`

message_body="$alert_header"

# Get the information for this Promise unit. Includes workaround for promiseutil tty issue.
screen -D -m sh -c "promiseutil -C subsys -v >$unit_ID_tmp"

# Drop the output into a variable.
unit_ID=$(<$unit_ID_tmp)

# Get the report, put it into a tmp file.
screen -D -m sh -c "promiseutil -C enclosure >$enclosure_results_tmp"

if ! grep -qv "Everything is OK" $enclosure_results_tmp
then
        enclosure_pass="false"
        # Get a more detailed report, put it into a tmp file.
        screen -D -m sh -c "promiseutil -C enclosure -v >$enclosure_results_tmp"

        # Build the message.
        message_body=$message_body$fail_header$unit_ID$(<$enclosure_results_tmp)
fi

#  ----------------- Logging & email ------------------

# Log the results, conditionally send email on failure.
if [ "$enclosure_pass" == "false" ]; then
        message_body="$message_body\n\n$alert_footer"
        echo "$DATESTAMP: \n\n$message_body" >> /var/log/system.log
        if [ "$send_email_alert" == "true" ] ; then
                "$sendemail_path" -f $alert_sender -t $alert_recipient -u $alert_subject -m "$message_body" -s $alert_smtp_server
        fi
else
        echo "$DATESTAMP: $pass_msg" >> /var/log/system.log
fi
# Cleanup
rm -f rm -f $unit_ID_tmp $enclosure_results_tmp

The script was developed against a Promise Pegasus2. It hasn’t been tested with the earlier Promise Pegasus series.

2014-11-07 – Update: Merci to Stéphane Allain for catching a typo in the script.

Promise Pegasus2: Scripting a disk check with promise_disk_check.sh

When you deploy a Promise Pegasus2, you want to run regular disk health checks and send an email notification if there’s a problem. The Promise Utility app can theoretically do this* when there’s someone logged in at the console, but we’re rarely running these in environments where there’s anyone logged at the console.

The solution is to script a check of the disks using the promiseutil command line utility and then create a cronjob to run it at regular intervals.

Here’s an example disk check that parses the output of phydrv, logs each run to system.log and can optionally send email if a problem is found.

#!/bin/bash
#
# promise_disk_check.sh
#
# Checks the phydrv status of a Promise Pegasus, logs and mails the output if there's an issue.
#
# Author: A @ Modest Industries
# Last update: 2014-07-19
# 2014-07-19 - tweaked grep to allow for Media Patrol
#
# Works with Promise Utility for Pegasus2 v3.18.0000.18 (http://www.promise.com)
# Requires sendemail for email alerts (http://caspian.dotconf.net/menu/Software/SendEmail/)

export DATESTAMP=`date +%Y-%m-%d\ %H:%M:%S`

# Editable variables

# Path to sendemail
sendemail_path="/Library/Scripts/Monitoring/sendemail"
# Email alert?
send_email_alert=true

# Variables for sendemail
# Sender's address
alert_sender="[email protected]"

# Recipient's addresses, comma separated.
#alert_recipient='[email protected], [email protected]'
alert_recipient="[email protected]"

# SMTP server to send the messages through
# alert_smtp_server="smtp.example.com:port"
alert_smtp_server="smtp.example.com"

# Subject line of the alert.
alert_subject="Alert: Promise disk problem detected on $HOSTNAME."

# Header line at the top of the alert message 
alert_header="At $DATESTAMP, a problem was detected on this device:\n"

# Pass / Fail messages
pass_msg="Promise disk check successful."
fail_msg=" *** Promise disk check FAILED!!! ***"

# ------------ Do not edit below this line ------------------
# Variables
pass=true
results=""

# Create temp files
unit_ID_tmp=`mktemp "/tmp/$$_ID.XXXX"`
results_tmp=`mktemp "/tmp/$$_results.XXXX"`

# Get header information for this Promise unit. Includes workaround for promiseutil tty issue.
screen -D -m sh -c "promiseutil -C subsys -v >$tmpdir$unit_ID_tmp"
unit_ID=$(<$tmpdir$unit_ID_tmp)

# Get status of the disks.  Includes workaround for promiseutil tty issue.
screen -D -m sh -c "promiseutil -C phydrv >$tmpdir$results_tmp"

# Check each line of the output the test results.
while read -r line
do
        if grep '^[0-9]' <<< "$line" | grep -Eqv 'OK|Media'
        then
                results=$results"BAD DRIVE DETECTED: $line\n\n"
                pass=false
        fi
done < $tmpdir$results_tmp

# Log the results, conditionally send email on failure.
if [ "$pass" = false ] ; then
        results="$alert_header$unit_ID\n\n$results\n$alert_footer"
        echo "$DATESTAMP: $fail_msg\n\n$results" >> /var/log/system.log
        if [ "$send_email_alert" = true ] ; then
                "$sendemail_path" -f $alert_sen:der -t $alert_recipient -u $alert_subject -m "$results" -s $alert_smtp_server
        fi
else
        echo "$DATESTAMP: $pass_msg" >> /var/log/system.log
fi

# Cleanup
rm -f $tmpdir$unit_ID_tmp $tmpdir$results_tmp

Note that the script uses sendemail for sending mail, a very useful little drop in for when the local machine isn’t running mail services.

*I say “theoretically” because configuring email in the Promise Utility is a mess and I’ve yet to see a single successful notification after configuring it.

2014-05-04 – Updated to make the path to sendemail a variable.

2014-07-19 – Changed grep to handle false positive during Media Patrol runs

Mac OS X automated backups of Server 2.x

My colleague and I had been trying to come up with a decent backup script for Server 2.x services on Mac OS X 10.8 Mountain Lion. We had the skeleton of a Bash script from another colleague but wanted more functionality, so we began bouncing the script back and forth between us. Before we knew it, we had a working system that we now both deploy on 10.8 Server systems, osx-backup.sh.

I like to keep things simple when it comes to backup and this script is no exception. Edit the variables at the top to suit your fancy and then create a nightly cronjob to run it as root and you’re done.

Among the services that get backed up are:

  • Open Directory archive
  • Profile Manager database
  • DNS records
  • Service plists
  • Wiki
  • CalDAV & CardDAV databases

The script automatically shuts down each service that needs to be disabled before backup, then re-enables the service as soon as possible.

On the weekend, over a big pile of bacon on the weekend, we decided to turn the script loose to the community, so my colleague has posted the script on his GitHub repository. This version works with Server 2.2.2 on 10.8.5.

Caveats: Like any backup system, read the script completely and test it before putting it into production.

Promise Pegasus2: The gap between a failing disk and a failed disk.

We were recently called in to diagnose a relatively new Promise Pegasus2 R6 that intermittently refused to mount. The Promise Utility app reported nothing amiss with the RAID or the drives, green lights everywhere, so we used the command line to dig a little deeper.

So let’s run a verbose SMART check on the unit:

promiseutil -C smart -v

The first three drives checked out. Drive 4 indicated that SMART thought everything was fine:

PdId: 4
Model Number: TOSHIBA DT01ACA2
Drive Type: SATA
SMART Status: Enable
SMART Health Status: OK

But then a little further down,  CRC errors:

Error 165 occurred at disk power-on lifetime: 1176 hours (49 days + 0 hours)
  When the command that caused the error occurred,
  the device was active or idle.

  After command completion occurred, registers were:
  ER ST SC SN CL CH DH
  -- -- -- -- -- -- --
  84 51 50 b0 ee 81 0d

  Commands leading to the command that caused the error were:
  CR FR SC SN CL CH DH DC   Powered_Up_Time  Command/Feature_Name
  -- -- -- -- -- -- -- --  ----------------  --------------------
  61 80 a8 80 ee 81 40 00      18:38:48.276  WRITE FPDMA QUEUED
  61 80 a0 00 ee 81 40 00      18:38:48.276  WRITE FPDMA QUEUED
  61 80 98 80 ed 81 40 00      18:38:48.276  WRITE FPDMA QUEUED
  61 80 90 00 ed 81 40 00      18:38:48.276  WRITE FPDMA QUEUED
  61 80 88 80 ec 81 40 00      18:38:48.275  WRITE FPDMA QUEUED

Error 164 occurred at disk power-on lifetime: 1175 hours (48 days + 23 hours)
  When the command that caused the error occurred,
  the device was active or idle.

  After command completion occurred, registers were:
  ER ST SC SN CL CH DH
  -- -- -- -- -- -- --
  84 51 10 f0 ad 6b 0d  Error: ICRC, ABRT 16 sectors at LBA = 0x0d6badf0 = 225160688

  Commands leading to the command that caused the error were:
  CR FR SC SN CL CH DH DC   Powered_Up_Time  Command/Feature_Name
  -- -- -- -- -- -- -- --  ----------------  --------------------
  35 00 80 80 ad 6b 40 00      18:36:07.145  WRITE DMA EXT
  35 00 80 00 ae 6b 40 00      18:36:07.144  WRITE DMA EXT
  35 00 80 00 ad 6b 40 00      18:36:07.144  WRITE DMA EXT
  35 00 80 80 ab 6b 40 00      18:36:07.139  WRITE DMA EXT
  35 00 80 00 ab 6b 40 00      18:36:07.139  WRITE DMA EXT

Error 163 occurred at disk power-on lifetime: 1175 hours (48 days + 23 hours)
  When the command that caused the error occurred,
  the device was active or idle.

  After command completion occurred, registers were:
  ER ST SC SN CL CH DH
  -- -- -- -- -- -- --
  84 51 f0 10 5e 5d 0d  Error: ICRC, ABRT 240 sectors at LBA = 0x0d5d5e10 = 224222736
...

The client confirmed that he’d seen a warning light on drive 4, but that it had “gone away”. We had them back the data up immediately. Promise support subsequently verified that the drive had failed based on the logs and sent a replacement drive out.

If the drive had failed completely, I assume the RAID would have kicked in, taken the bad drive offline and continued spinning, but since the drive hadn’t actually failed, the volume was struggling with a failing member and that was causing boot and performance issues.

The take-away is that there’s a generous gap between a drive that’s beginning to fail and a drive that’s failed enough for the Promise Utility app to detect it. Verbose mode is your friend.

Turn off Promise Pegasus2 power savings one-liner.

From Terminal:

promiseutil -C ctrl -a mod -s “powersavinglevel=0″

Now check your handiwork:

promiseutil -C ctrl -v | grep PowerSavingLevel

You should see:

PowerSavingLevel: 0

You may be tempted to turn off PowerManagement. Here’s what Promise’s website has to say about that:

Do not disable Power Management. Doing so will make the Pegasus2 hang when the Mac is in sleep or shut down and restarted and will require the Pegasus and the Mac to be power cycled to return to normal operation.

Promise Pegasus2 command line tools.

When I began deploying Promise Pegasus2 storage devices, I wasn’t happy with the state of the Promise Utility app. It doesn’t provide email alerts except when a user is logged in and this is isn’t optimal for most of our deployments.

Then I stumbled on a couple of Ruby scripts by GriffithStudio that showed a way around many of the limitations of the Promise GUI.

When you install the Promise Utility for Promise Pegasus2, it includes a command line utility. You can view status and even change the settings of the device. In a Terminal window, type:

promiseutil

You’ll be greeted with an interactive command line.

-------------------------------------------------------------
Promise Utility
Version: 3.18.0000.18 Build Date: Oct 29, 2013
-------------------------------------------------------------
 
List available RAID HBAs and Subsystems
=============================================================
Type  #    Model         Alias                            WWN                 
=============================================================
hba   1  * Pegasus2 R4                                    2000-0001-5557-98bf 
 
Totally 1 HBA(s) and 0 Subsystem(s)
 
-------------------------------------------------------------
The row with '*' sign refers the current working HBA/Subsystem path
To change the current HBA/Subsystem path, you may use the following command:
  
  spath -a chgpath -t hba|subsys -p <path #>.
 
Type help or ? to display all the available commands
-------------------------------------------------------------
 
cliib> 

To get a list of commands, type ? and press return. Some of the available commands include:

subsys - Model, serial number, hardware revision.
enclosure - Enclosure status.
ctrl - Firmware version, array & RAID status.
phydrv - Physical drive status.
array - Array status.
logdrv - Logical drive status.
event - Event log, including abnormal shutdowns.

Many of the commands yield a brief Pass/Fail style response:

cliib> enclosure 
=============================================================
Id  EnclosureType               OpStatus  StatusDescription                    
=============================================================
1   Pegasus2-R4                 OK        Everything is OK

If you want more details, you can add the verbose flag, -v. Want the serial number, model and hardware revision?

cliib> subsys -v
 
-------------------------------------------------------------
Alias: 
Vendor: Promise Technology,Inc.        Model: Pegasus2 R4
PartNo: F29DS4722000000                SerialNo: M00H00CXXXXXXXX
Rev: B3                                WWN: 2000-0001-5557-98bf

You can grab enclosure information, including temperature of box, backplane and controller card, as well as the rotation speed of the fans and voltage on the power rails with this:

cliib> enclosure -v
 
 
-------------------------------------------------------------
Enclosure Setting:
 
EnclosureId: 1
CtrlWarningTempThreshold: 63C/145F     CtrlCriticalTempThreshold: 68C/154F
 
 
-------------------------------------------------------------
Enclosure Info and Status:
 
EnclosureId: 1
EnclosureType: Pegasus2-R4
SEPFwVersion: 1.00
MaxNumOfControllers: 1                 MaxNumOfPhyDrvSlots: 4
MaxNumOfFans: 1                        MaxNumOfBlowers: 0
MaxNumOfTempSensors: 2                 MaxNumOfPSUs: 1
MaxNumOfBatteries: 0                   MaxNumOfVoltageSensors: 3
 
=============================================================
PSU       Status                        
=============================================================
1         Powered On and Functional     
 
=============================================================
Fan Location        FanStatus             HealthyThreshold  CurrentFanSpeed 
=============================================================
1   Backplane       Functional            > 1000 RPM        1200 RPM        
 
=============================================================
TemperatureSensor   Location       HealthThreshold   CurrentTemp    Status    
=============================================================
1                   Controller     < 63C/145F        49C/120F       normal    
2                   Backplane      < 53C/127F        47C/116F       normal    
 
=============================================================
VoltageSensor  Type    HealthyThreshold         CurrentVoltage  Status         
=============================================================
1              3.3V    +/- 5% (3.13 - 3.46) V   3.2V            Operational    
2              5.0V    +/- 5% (4.75 - 5.25) V   5.0V            Operational    
3              12.0V   +/- 8% (11.04 - 12.96) V 12.0V           Operational 

How about the state of the physical drives?

cliib> phydrv   
=============================================================
PdId Model        Type      Capacity  Location      OpStatus  ConfigStatus     
=============================================================
1    TOSHIBA DT01 SATA HDD  2TB       Encl1 Slot1   OK        Array0 No.0      
2    TOSHIBA DT01 SATA HDD  2TB       Encl1 Slot2   OK        Array0 No.1      
3    TOSHIBA DT01 SATA HDD  2TB       Encl1 Slot3   OK        Array0 No.2      
4    TOSHIBA DT01 SATA HDD  2TB       Encl1 Slot4   OK        Array0 No.3

You can also run commands without entering interactive mode and this is useful when incorporating into bash scripts. Simply add the -C flag, followed by the command you want to run. For example:

krieger:~ admin$ promiseutil -C logdrv -v

will let you view the logical drives.

Many of these commands will run on the previous generation Promise Pegasus units.

Be aware: it’s possible to change the configuration of your Pegasus2 or even destroy your RAID setup from the command line, so use caution when working on production systems.