2

I am currently running CentOS 5 and looking for a terminal command that can allow me to monitor the status of the RAID set up (ie, if a drive is down) without having to get into the kernel. This is a live web server afterall.

update: specs are its a dell sc1435 with SAS 5i/R controller.

2
  • no matter what raid you have [ software / hadrdware ] remember to periodically run consistency check. and to have monitoring system to alert you when first disk fails
    – pQd
    Jun 16, 2009 at 20:00
  • Please include specific information about what kind of RAID setup you have. Is it md (software RAID), hardware? If it's hardware, what controller/driver?
    – freiheit
    Jun 16, 2009 at 20:14

6 Answers 6

6

If you're using software RAID with a regular disk controller then use:

mdadm --detail <dev>

where is /dev/md0 for example. This will show you the current status. If a drive fails, you'll also see lots of nastiness in /var/log/messages.

5

it's raid dependent. for lsi [ it's in plenty of dell and hp servers ] you use tool called MegaCLI.

for 3ware cards - tw_cli

it usually comes with 'drivers' or documentation for your hardware.

4

If this is a software raid (mdadm) and you want to look at the current status you could simply do a cat /proc/mdstat. If you want to have something where the screen is refreshed every 10 seconds you could do a watch -n 10 cat /proc/mdstat.

3

Dell probably provide a tool to monitor it, but I can hazard a guess that it will be bloated and implemented with Java, like most lacklustre OEM utilities.

Fortunately enough it appears that the SC1435 is supported by the wonderful mpt-status utility. Just make sure that you have the following options enabled in your kernel:

CONFIG_FUSION=y
CONFIG_FUSION_SAS=y
CONFIG_FUSION_MAX_SGE=128
CONFIG_FUSION_CTL=y

You can then use mpt-status from the CLI to view the health of your RAID array.

I personally then use a simple Python script called from cron which periodically checks the status and issues us alerts by email. In a similar fashion to how mdadm behaves. You can of course specify how frequently you wish to check. Feel free to use it yourself:

#!/usr/bin/env python

# Copyright (c) 2009 Dan Carley <dan.carley@gmail.com>
#
# Permission to use, copy, modify, and distribute this software for any
# purpose with or without fee is hereby granted, provided that the above
# copyright notice and this permission notice appear in all copies.
#
# THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
# WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
# MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
# ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
# ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
# OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

"""
Report failures from hardware RAID controllers.

Requires the supporting utilities:
    mpt-status(8)   for MPT controllers.
    tw_cli(8)       for 3ware controllers.

Intended to be scheduled from crontab as follows:
    MAILTO="alert@your.domain"
    0   */3 *   *   *   /usr/local/sbin/hwraid_monitor.py options
"""

from re import search
from sys import exit
from os.path import isfile
from optparse import OptionParser
from subprocess import Popen, PIPE

def check_controller(type):
    ret = True

    if type == 'mpt':
        cmd = [ '/usr/sbin/mpt-status', '-s' ]
        array = {'regex': '^log_id$',
                 'pos': 2,
                 'string': 'OPTIMAL'}
        drive = {'regex': '^phys_id$',
                 'pos': 2,
                 'string': 'ONLINE'}
    elif type == 'tw':
        cmd = [ '/sbin/tw_cli', 'info' ]
        contr = {'regex': '^c\d+$'}
        array = {'regex': '^u\d+$',
                 'pos': 2,
                 'string': 'OK'}
        drive = {'regex': '^p\d+$',
                 'pos': 1,
                 'string': 'OK'}

    if not isfile(cmd[0]):
        print "%s: Utility not found" % cmd[0]
        return False

    if type == 'tw':
        controllers = []
        p = Popen(cmd, stdout=PIPE)
        o, e = p.communicate()
        if e:
            print e
        for c in o.split('\n'):
            c = c.split()
            if len(c) > 2 and search(contr['regex'], c[0]):
                controllers.append(c[0])
    elif type == 'mpt':
        controllers = ['']

    for c in controllers:
        p = Popen(cmd + [c], stdout=PIPE)
        o, e = p.communicate()
        if e:
            print e.split('\n')
        for v in o.split('\n'):
            v = v.split()
            if len(v) > 2:
                # Array check.
                if search(array['regex'], v[0]) and v[array['pos']] != array['string']:
                    print "Array failure: \n\t%s" % '\t'.join(v)
                    ret = False
                # Drive check.
                if search(drive['regex'], v[0]) and v[drive['pos']] != drive['string']:
                    print "Drive failure: \n\t%s" % '\t'.join(v)
                    ret = False

    return ret

def main():
    usage = "usage: %prog options"
    parser = OptionParser(usage=usage)
    parser.add_option("--mpt", action="store_true", default=False,
                      dest="mpt", help="MPT controller support.")
    parser.add_option("--tw", action="store_true", default=False,
                      dest="tw", help="3ware controller support.")
    (options, args) = parser.parse_args()

    if not options.mpt and not options.tw:
        parser.print_help()
        exit(2)

    fail = False

    if options.mpt:
        if not check_controller('mpt'):
            fail = True

    if options.tw:
        if not check_controller('tw'):
            fail = True

    if fail:
        exit(1)

if __name__ == "__main__":
    main()
0

mdadm offers detailed information about every aspect of the Linux software raid.

0

watch -n 10 cat /proc/mdstat this will setup on u r system

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.