mdadm --monitor does not work in any way. mdadm --monitor does not mail or run a program on status changes in md-arrays. No difference if started manually or by /etc/init.d/mdmonitor Neither the MAILADDR nor PROGRAM option in mdadm.conf work. Reproducibility Every time. Steps to reproduce 1. Make md-array 2. add MAILADDR root to /etc/mdadm.conf 3. /etc/init.d/mdmonitor start 4. offline one disk in md-array. Actual results Nothing. Expected Results Mail should be sent alerting admin to status change.
Last tested configuration (have tested several): --------------------------------- $ cat /proc/mdstat Personalities : [raid1] read_ahead 1024 sectors md0 : active raid1 hde1[0] hdg1[1] 180939968 blocks [2/2] [UU] unused devices: <none> ------------------------------ Example of failure not reported by mdadm --monitor : ------------------------------ Oct 17 13:23:23 xxx smartd: Device: /dev/hdg, S.M.A.R.T. Attribute: 1 Changed -2 Oct 17 13:53:23 xxx smartd: Device: /dev/hde, S.M.A.R.T. Attribute: 1 Changed 1 Oct 17 13:53:24 xxx smartd: Device: /dev/hdg, S.M.A.R.T. Attribute: 1 Changed 2 Oct 17 13:58:32 xxx kernel: hde: dma_intr: status=0x51 { DriveReady SeekComplete Error } Oct 17 13:58:32 xxx kernel: hde: dma_intr: error=0x40 { UncorrectableError }, LBAsect=10458, high=0, low=10458, sector=10392 Oct 17 13:58:32 xxx kernel: end_request: I/O error, dev 21:01 (hde), sector 10392 Oct 17 13:58:32 xxx kernel: raid1: Disk failure on hde1, disabling device. Oct 17 13:58:32 xxx kernel: ^IOperation continuing on 1 devices Oct 17 13:58:32 xxx kernel: raid1: hde1: rescheduling block 10392 Oct 17 13:58:32 xxx kernel: md: updating md0 RAID superblock on device Oct 17 13:58:32 xxx kernel: md: (skipping faulty hde1 ) Oct 17 13:58:32 xxx kernel: md: hdg1 [events: 0000004a]<6>(write) hdg1's sb offset: 180939968 Oct 17 13:58:32 xxx kernel: md: recovery thread got woken up ... Oct 17 13:58:32 xxx kernel: md0: no spare disk to reconstruct array! -- continuing in degraded mode Oct 17 13:58:32 xxx kernel: raid1: hdg1: redirecting sector 10392 to another mirror
When will this be fixed? Three open bugs for mdadm for two month... thats not quite the enterprise support I expect...
Christian Hofmann: Bugzilla is not a support mechanism. If you need or expect support you should use RH support not bugzilla.
Christian - did you end up calling support? Do you need more information on how to do that?
mdadm --monitor is working fine here. Current version is 1.4.0-1 so make sure you are up to date with current version and then make sure that you are running some sort of smtpdaemon so that the program can successfully email the person listed in the MAILADDR line.
Closing this bug since mdadm --monitor works in recent versions.