[Info-vax] How do I avoid a full disk shadow copy when I shutdown/boot a node back into the cluster?

[Phillip Helbig---undress to reply]

In article
<1281120.105.1333567428620.JavaMail.geo-discussion-forums at pbnt10>, Ken
Fairfield <ken.fairfield at gmail.com> writes: 

> 1) Be sure you've configured and enabled HBMM (Host-Based Mini-Merge) 
>    on all your shadow volumes.

In his case (one node and its shadowset members go away and come back), 
isn't MINICOPY more relevant?

Another idea: I have this set up:

   SHADOW_MBR_TMO=300        ! allow one node to reboot

The idea is that if the reboot takes less than 5 minutes, no copy or 
merge will be needed, but of course I/O will hang during that time.  
Whether or not this is the better solution than MINICOPY/MINIMERGE 
depends on your applications.  It's a DYNAMIC parameter, so you could 
have a default value for unplanned reboots and perhaps change it 
temporarily for planned reboots.