When you pull out a disk from your Raid setup it shows a warning in syslog

1. Jan 27 10:18:22 EL860 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0019): Drive
2. removed:port=1.
3. Jan 27 10:18:22 EL860 kernel: 3w-9xxx: scsi0: AEN: ERROR (0x04:0x0002): Degraded
4. unit:unit=0, port=1.

However if no one is looking at syslog that won’t really be helpfull.

3Ware provides a tool from their site called tw_cli which can be used to manage
the raid setup from the command line.

1. [EL860-root@EL860 admin]# tw_cli /c0 show

3. Unit  UnitType  Status         %RCmpl  %V/I/M  Stripe  Size(GB)  Cache  AVrfy
4. ------------------------------------------------------------------------------
5. u0    RAID-1    REBUILDING     41%     -       -       232.82    RiW    ON

7. VPort Status         Unit Size      Type  Phy Encl-Slot    Model
8. ------------------------------------------------------------------------------
9. p0    OK             u0   232.88 GB SATA  0   -            ST3250310NS
10. p1    DEGRADED       u0   232.88 GB SATA  1   -            ST3250310NS

I’d figure I’d either have to write wrapper script around that or find some other way of integrating it.
Asking the question on ##infra-talk on irc.freenode.net gave me the following link to a check script on github

koollman: sdog: something like http://github.com/stanaka/check_tw should work.

With that in your snmpd.conf you can get the info via snmp

1. [root snmp]#  snmpwalk  localhost   -v 2c  -c public  .1.3.6.1.4.1.2
2. 021   | grep ext
3. UCD-SNMP-MIB::extIndex.1 = INTEGER: 1
4. UCD-SNMP-MIB::extNames.1 = STRING: TW_RAID
5. UCD-SNMP-MIB::extCommand.1 = STRING: /usr/local/sbin/check_tw
6. UCD-SNMP-MIB::extResult.1 = INTEGER: 2
7. UCD-SNMP-MIB::extOutput.1 = STRING: CRITICAL: Unit: u0, Type: RAID-1, Status: RE
8. BUILDING
9. UCD-SNMP-MIB::extErrFix.1 = INTEGER: 0
10. UCD-SNMP-MIB::extErrFixCmd.1 = STRING:
11. UCD-SNMP-MIB::ssSysContext.0 = INTEGER: 2073
12. UCD-SNMP-MIB::ssRawContexts.0 = Counter32: 11781783
13. UCD-DLMOD-MIB::dlmodNextIndex.0 = INTEGER: 1