Well I finally finished the latest update to the SRDF SRA TechBook.
It now includes all SRAs up to, and including, the most recent release 6.1. You’ll see this mentioned in the preface to remind you of that:
- SRM 6.1 and stretched storage
- VMAX All Flash
- SRDF Adapter Utilities (SRDF-AU)
That last bullet point is probably the one that customers reference most in the TechBook since it shows what SRA versions are compatible with what VMAX storage. This is the latest one below. Note that it contains a version that was not released with any fanfare so many customers missed it. It is version 6.0.1. For all intents and purposes, it is the same as version 6.0 but it adds support for our VMAX All Flash arrays so that customers do not have to upgrade to 6.x if they are running SRM 5.5.
I tried to include as much information on the relatively new SRDF Adapter Utilities as possible without increasing the book length significantly. It’s still under 500 pages and I want to keep it that way. For instance, when configuring test failover you can use a variety of TimeFinder technologies. I only included the SRDF-AU in one of them rather than create duplicate material. In the TB I keep to using the VSI SRA-U as the foundation utilities for most examples because it is still heavily utilized by customers and the screens are so different from the SRDF-AU that it would be too confusing to drop it altogether. Maybe in the next version, we’ll see.
You can check out the latest version here: SRM TB.
On a related note, I wanted to mention an issue that I found in my SRDF/Metro SRM testing that may be of interest. This issue is in fact a problem in the VMware SRM code that impacts a particular use case involving disaster recovery in SRDF/Metro. It isn’t, however, related to the other issues I’ve mentioned around SRM and stretched storage (SRDF/Metro with SRM). Anyway the issue will only impact you if you run a disaster recovery while the Protected Site is still operational from an EMC storage perspective – so the array manager and SRM are operational though of course your applications could be down. Some customers I know have a requirement to run a DR test rather than a test failover, and they do so with the Protected Site operational, so this would impact them. Fortunately this problem can be avoided with a couple of simple changes.
First, it will be necessary to change an advanced SRM parameter named recovery.forceRecovery on the Recovery Site. By default this parameter is set to false. To change this, highlight the Recovery Site, select the Manage tab and Advanced Settings sub-tab. Then select Edit and check the box to enable the force recovery.
Once it is enabled, DR can be run. The parameter will enable an additional checkbox in the DR dialog called “Forced recovery – recovery site operation“. This box must be checked when running the DR.
When the DR run completes, although it will indicate success, another DR run will be needed. Here is the result you will see:
Simply execute the DR again. This time there will be no radio buttons or check boxes.
The final result will show a successful recovery.
Remember this only affects SRDF/Metro environments . If you want the inside scoop on the expected errors, etc., see the following KB: https://support.emc.com/kb/483401.