At the moment I’m working through the typo corrections for the second pressing of the SRM 5.0 book. In the main these are little corrections in the text. But occasionally, I come across something a little bit more major. You would not believe how much work goes into the review process – but even the most rigorous process – its still possible for something to slip through the net. I guess its testament to how reading something on screen is never the same as reading in the printed hard-copy format.
There will be an official errata but in the interest in turning these around rapidly I wanted to blog about them here.
I intend to main this page as find errors or mistakes. If you spot any technical errors or mistakes like this – let me know and if I think they warrant it I will include them on this page…
iSCSI Chapters and Static Discovery
There’s a couple of chapters where I cover iSCSI in EMC Celerra, Dell Equallogic and HP P4000 – where I state that the iSCSI Software Initiator does not support “static discovery” of target. Actually, if you look into the static discovery tab after using a iSCSI SRA you will see entries populated there.
page 91 – At the Protected Site – (With thanks to Michael Armstrong)
This typo was found my Michael Armstrong. In it I refer to carrying at task out on a EMC Clarrion at the Protected Site. Step 1 however, erroneously says that “in my case, this called “New_Jersey_Cluster1″. That should, of course read “New_York_Cluster1″. As a consequence the graphic (Figure 4.22) is incorrect as it shows me selecting the “New Jersey” storage group in UniSphere, when it should have the “New York” storage group selected instead.
page 209 – vSphere Replication and IP Pools
In the beta programme the set-up of VR required the use of IP Pools. This was subsequently dropped in the release candidate and the GA. This reference slipped through the net. To be clear, VR does NOT require IP Pools.
page 218 – NFS and iSCSI
In the VR chapter I have a screen grab where a VM on local storage in the Protected Site being replicated to local storage in the Recovery Site. However in the text I say that one VM is on NFS storage, and the other is on iSCSI storage. That isn’t the case. It’s not a showstopper – if you remember VR does not care what the datastore type is…
page 228 – SRM Communication on port 80
My explanation of why the SRM installer uses Port 80 in the book isn’t a bad one. But after publication I found a better explanation which I blogged about recently. It’s largely of acedemic interest this one, but in the interests of being as technically accurate this better…
page 237 – Invalid Site name mentioned
This is a cosmetic error. Throughout the book I make reference to site of New York, New Jersey and later still Washington DC. For some bizarre reason I name check “Chicago” here. That should really read “New Jersey”.
pg 265 – Recovery Plans and Connected CD-ROMs
This is not so much an error, but a failure of mine. Looking back I think I could have probably explained the source of this problem using less words! All I was trying to say was – as with VMotion, SRM can get unhappy with CD-ROMs left connected to the VM. The example I gave was where an error in a VUM update left the VMwareTools .iso connected to a VM I was trying to recover. In hindsight I think I would recommend using PowerCLI on a regular basis to check for connected CD-ROMs and then running a script to disconnect them. Sorry for getting lost in the detail!
pg 306 – How to do addy ups and take-aways…
Here there is slight miscalculation. I talk about although you might have 10 VMs that make up an application, you might decide to recovery all 10, but only power on 7 – because 7 VMs is enough to make the application meet its minimum QoS demands. Oddly, I precede to say that this will lead 2 VMs left over if one of the other 7 fail. I do, of course know that 10-7 = 3, not 2. What can I else can I say?
pg 342 – Ch-ch-changes in the Recovery Site
Bit of poor phraseology here I think. All I’m trying to say is now that we have looked at how changes in the Protected Site affect your SRM configuration, I’m going to look at changes in the Recovery Site to outline the impact there.
pg 461 – RDMs can be greater than 2TB in size…
This is a real howler. It is possible in vSphere5 to have RDMS that greater than 2TB in size. Remember you do need the right sort of RDM to allow the guest operating system to be able to use is own GPT framework – you need a pass-through RDM…