MCS7835-H1 vs. MCS7835-H2 (aka U320 vs. SAS)

One of many steps involved in moving to our new building is the implementation of Cisco IP telephony. To support that implementation, it was necessary to relocate our existing voice servers to our colocation site (along with their associated VLAN). This would pave the way for installing additional servers in the coming weeks.

To prepare for this migration, we performed some preliminary testing last weekend. During our maintenance window, we shutdown the network interfaces to the Cisco voice servers and verified that the IP phones registered with an alternate subscriber in our Houston office. Everything worked as planned. So this weekend we were scheduled to perform the relocation from 10PM Saturday through 2AM Sunday.

We shut the servers down, removed them from the rack, and transported everything to the colocation site. The voice VLAN was migrated, and the Call Manager Publisher and Subscriber came up without issue. But the Unity server (an MCS7835-H1) was not booting. It would power on, but there was no video signal and the hard drives never spun up. My diagnosis – system board failure (the video and the drive controller are both integrated on the system board).

We opened a P1 ticket with Cisco TAC around 3AM, and an RMA was issued. We have a four-hour response agreement in place on all of our Cisco products. Around 5AM, logistics emailed to advise that the replacement server (and a field engineer) would not be out until Tuesday morning. Not acceptable. We called logistics back and a new RMA was entered. They said the part was not stocked at the local depot, and that it would arrive on the “next flight”. By around 7AM we had confirmation that the server would arrive at 9:30AM and a tech would be about an hour behind.

The server actually showed up closer to 8:30AM (and it did come from a local parts depot by the way). I unboxed it, and it was the wrong server. They sent us an H2 instead of an H1. Why would this matter? They use different types of hard drives. We wanted to transfer the drives (with our data) to the replacement server and boot up. Not possible with an H2. So back on the phone with logistics. Logistics sent me back to TAC to get a new RMA for the correct server type. TAC said they wanted to get HP on the line to see if we could just get replacement parts for our H1 instead of an entire replacement server.

Meanwhile, a tech showed up around 10AM from the original call (this was the tech that was supposed to be dispatched after the server arrived). At the same time, a voice engineer from General Datatech was kind enough to come out and help us. Since the hardware tech that came out was having no more luck with the dead server than we were, the engineer and I decided to proceed down a different path. We opted to shutdown the subscriber and boot the Unity hard drives from that server. The phones registered to the publisher, and the Unity drives came up without issue. Voicemail was working again!

So one of my colleagues was kind enough to come stand in for the rest of the process. I finally left around 1PM and came home to sleep. Is there a point to this story? Not sure – its just an account of what I did for 15 hours this weekend. Would I have done anything differently? Not really. Hardware failures happen, and there is always danger in shutting down servers and relocating them. I’ve seen many a server in my career that was working fine until it was powered down. You don’t even have to move them to make them break. Anyway, my only regret was not more closely checking on the model number of the original RMA. I’m still not certain if that was our mistake or Cisco’s, and I’ll probably never know.

Let me close by thanking everyone again for their help and assistance this weekend. My colleagues and my Cisco VAR, General Datatech, really came through for us. There are many partners you can choose from, but my team at GDT is a real class act. Thank you all.

1 Comment »

  1. your Sister said,

    April 22, 2008 @ 2:55 pm

    Wow, sounds like a crazy weekend. I’m glad it all worked out. Take a nap!

RSS feed for comments on this post · TrackBack URI

Leave a Comment

Creative Commons License Creative Commons License