3

This concerns a server that runs ESXi (v. 4.x or 5.x) installed on drives that are configured into a raid10 using an LSI 3ware 97050 raid controller.

I would like to know if there is a way to monitor the LSI 3ware series of controllers, in particular the 9750, through ESXi. And to hopefully also run the monitoring daemon LSI provides.

I know you can set up a cronjob to execute tw_cli through ssh on the ESXi server. However that's not really ideal.

I am not using vcenter by the way. It would be nice to have more than just monitoring working, since the 3ware software has a very useful web client, besides tw_cli.

2
  • 1
    NOTE: One reason to use 3Ware controllers over the competition: 3Ware controllers have the monitor client written in Java. In theory it is cross platform and would work on any system and probably would/does .
    – djangofan
    Apr 12, 2012 at 22:04
  • Right, in addition the controllers work well under linux.
    – aseq
    Apr 12, 2012 at 23:03

2 Answers 2

3

vCenter has some pre-built alerts that will catch hardware failure on (most) hardware. This will show up as an alert in the vSphere Client, and can also send out a trap to a listening host.

9
  • Will it detect drive failures on the raid?
    – aseq
    Apr 12, 2012 at 1:34
  • 1
    @aseq Pull a drive from a raid set and find out? Seriously, you should have tested that before putting things into produciton. I know we did on our Dell boxes; and yes it does alert in vCenter. If your controler is on the vMware hardware compatability list for the version of ESXi you are running, it should probably work as well. But you do need to actually test that reality matches the vendor literature.
    – rmalayter
    Apr 12, 2012 at 2:42
  • The controller is not on the compatibility list. I had to create a custom ESXi CD with the driver as provided by LSI included. That's why I didn't expect it to work through vcenter or vsphere client. None of the solutions I found online talked about that either. The only solution I found was the cron job I mentioned. The 4.x version of ESXi didn't even boot from the raid controller.
    – aseq
    Apr 12, 2012 at 18:15
  • I'd like to strongly encourage you to choose a different RAID controller. ESXi is not the sort of system that you want to hack together. There is tons of value in knowing your hypervisor will work correctly. Apr 12, 2012 at 18:29
  • 1
    That comment was not helpful. I fail to see how this can be "hackish at best" when LSI develops and releases drivers specifically for ESXi.
    – aseq
    Apr 19, 2012 at 0:13
1

It looks like the only way to monitor this 3ware raid controller through ESXi is to enable ssh access on the ESXi server, install tw_cli on the datastore (not in places such as /opt because it won't survive a reboot) and set up a daily or weekly cron job to run something like:

ssh esxihost.example.org /vmfs/volumes/datastore1/3ware/tw_cli /c0 show all
ssh esxihost.example.org /vmfs/volumes/datastore1/3ware/tw_cli /c0 show alarms

And have the result emailed to you. Which cron should do automatically. You can enhance it with grep and whatever else works to only email you in case of an alarm. Make sure to copy your ssh key so you don't need to provide your password. You will have to repeat this after the ESXi server reboots.

This has the added benefit you can also manage the raid using the tw_cli tool through an ssh session.

ssh can be enabled on ESXi 5.0 rather easily through the vsphere client:

Or using the console:

To enable ssh on ESXi 4.1 try this:

And on ESXi 4.0 this will enable ssh:

Since for 4.0 it's a little bit more hassle I laid out the basic upshot:

  • at the console hit alt F1
  • type: unsupported
  • enter password
  • vi /etc/inetd.conf
  • comment in the line(s) with ssh
  • kill inetd
  • run inetd (just typing it)

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .