Marvell 88E8056 and ESXi 4.1

So I have an older development/experimental server that runs couple of VMs on ESXi 4.1. The server’s motherboard (ASUS P5BV-C/4L) is from an old workstation, and it has integrated quad NICs which would be nice to be able to use.. except that the default build of ESXi 4.1 doesn’t see them (even though ESXi 4.1 technically supports Marvell 88E8056 NICs).

There are several pages that discuss the issue extensively, and have a lot of good information on them. Yet another page has a quick low down on how to get the driver properly installed.

However, having not worked on ESXi CLI for some time I had forgotten, for example, that busybox that ESXi uses wipes the root files on every reboot. After a while I recalled (from an old note) that to save changes to the /etc/vmware/simple.map I would need to execute /sbin/backup.sh 0 /bootbank/ after making the edits. But even that was unecessary.

One sentence on the brief recap page, would have saved me couple of hours tonight. So here it is: »Just upload the attached oem.tgz into /bootbank folder with scp, then reboot, and you’re done!» And when you do that, you are done – the pre-prepared oem.tgz works perfectly!

Yes, had I known, I would’ve known, but I didn’t. 🙂 Hopefully this saves time for someone else!

How To Destroy Brand Confidence In One Hit

Couple of a weeks ago the power supply of one of our PCs failed. The system is about one and half years old, so fine – sometimes things break. I replaced the power supply and the system was back up and running. The failed OCZ power supply was under warranty, and so I sent it in for replacement. The usual way: I pay the shipping in, they pay the shipping back.

The replacement – a reconditioned unit – arrived yesterday. First thing I noticed that the “WARRANTY VOID IF BROKEN OR REMOVED” sticker was, well, broken. Hm. So today I went ahead and installed it back into the original system, taking out the temporary PSU. I really hate replacing power supplies when the case is even slightly congested and this one was pretty tough to get to. Finally, the replacement PSU was in place and I hit the power button. Nothing!

Few moments of testing later I had determined that the unit OCZ sent as a replacement was DOA. I had to rip it out and put the temp PSU back in. Note to self: from now on the PSU replacement protocol will include stand-alone testing the new unit before I touch the target system.

I very much doubt the unit broke in transit; I’m guessing they either didn’t test it after repairs were completed, or for some reason the unit was never serviced and so I got someone else’s failed PSU in exchange to PSU I sent in for warranty service. The only way OCZ can salvage the situation at this point and avoid getting on my bad list is if they offer to pay shipping both ways. It’s not that much money, but the time wasted on this far exceeded what the replacement is worth.

Many companies forget that the value of warranty they offer not only comes from what they can advertise but also from the PR – positive or negative – depending how they handle warranty.

UPDATE 21 January 2010: Five Star Damage Control

OCZ handled the situation about the only way they could’ve handled it to minimize the negative impression that had already been created: they offered a free upgrade to a new product, or a pre-paid shipping label to return the DOA unit to service/exchange (apparently in case I had to keep the same model, such as a component for a tightly specified system). I chose the upgrade. Let’s hope the new unit works! 😉

UPDATE 22 January 2010: Not so fast, my friend

More to come. Dealing with OCZ tech support turned out to be less than what the first impression promised.