In my lab I’m using a Proliant DL380 G5 filled with disk as a storage server, running some VSA on top of the volume I created and formatted with a vmfs filesystem.
I was running into performance issues, storage was performing poorly, and I needed to check for usual problems like for example cache battery status, since on a P400 raid card, write back is by default disabled if the battery has a low charge, thus lowering performances to a minimum.
My problem however was, I was running on top of this server all my iscsi shared storage, so it was not so easy to poweroff the server to open up the raid card Bios anche check. Also, every reboot cycle is really time consuming. So, since HP has always had management software for their raid controllers available for windows or linux, I went to check if there was some possibilities also for VMware ESXi.
First, I checked on HP website and there was some promising download:
Look, a VIB file! 🙂
I downloaded it and installed in the usual way on my ESXi 5.1 server (even if the package is officially listed for ESXi 5.0…).
I also installed the utilities to manage the several HP components, since the first one was only the driver. If you want to install only the Raid utility, look for the hpacucli vib file.
And I finally reboot the server in order to complete the installation, hopefully for the last time…
After reboot, I checked the components were installed correctly:
From here, let the fun begin! There are many commands you can run on the raid controller by running /opt/hp/hpacucli/bin/hpacucli, I’m going to show you some of them:
First, I identified the controller. Remember is in slot 3, this will be needed in the next commands.
As I said, my problem seemed to be related to the cache battery, so I went to check its status:
The status seems to be ok, but I went to enable anyway the cache, the command is:
ctrl slot=3 modify dwc=enable forced
there are many commands as I said, and I found this page with many example. Go to check it. As a last example, a fun stuff:
First, I identified the physical disks in the controller (pd), then I enabled the led on every single disk. This is the result 🙂
excellent thank u very much
Very useful, thanks!
Just a head’s up, the RAID management utility has been renamed from hpacucli to hpssacli on newer Proliant servers.
Nice article! Really helped me out. This all will work fine as long as the Write Back Cache is working properly. The battery on my RAID controller seemed to be dead :).
Excellent thank you
Again, the advice of someone who doesn’t know what he talks about. With that command, you activate the write cache ON THE DISKS, which isn’t protected by anything. It’s stupid to do so. What you want is active the write cache on the CONTROLLER, then configure the cache ratio if necessary. It’s sickening how many people give this bad and dangerous advice. The command you need is hpacucli ctrl slot=0 modify nbwc=enable to enable the write cache even if the battery is faulty or not present. It’s dangerous so don’t do it in a production environment unless absolutely necessary. Get a new battery ASAP. And stay away from the drive write cache (dwc).
Thx for your feedback 🙂
Is it working with ESXi 6.0 ?
Tried it yesterday: Yes it works for me with ESXi 6.0.0 on a DL180 Gen9.
But you have to use the hpssacli for newer Proliant generations an on ESXi 6 and the interactive mode does not work.
great post (even 4 years on), some things to add:
– the official HPE VMWare ISO image from hpe.com already contains the required vib files loaded, hence you dont need to add them if you are running the HPE VMWare build on your server (which you should be 😎 )
if you mess with the raid layout (e.g clearing it and setting it up to RAID5 ) you probably need to reboot the server before you continue as it appears that ESX only scans the disks on startup.
After installing on esxi 6.0 scsi-hpsa drivers, problems with controller started. Datastore turns off-on after 40 days of uptime… updated to vmware drivers back. be careful with drivers from hewlett for vmware