Question about the Cluster Configuration Example


#1

Hi,

I am reading the user guide of version 13.9 of SicDB, regarding to the Cluster Configuration Example I have a few questions:
1.The fist line says there are 4 identical servers in the cluser, but only two are defined in the configuration file, are the other two omit?

2.The Example define server-0 as “server-0=p4xen1.local.paradigm4.com,7”, is that mean besides the coordinator instance, there will be other 7 worker instances running on server-0.

3.Server 1 is defined as “server-1=p4xen2.local.paradigm4.com,8”, that is there will be 8 worker instances running on server-1, but the data-dir-prefix section only defines 4 entries:
data-dir-prefix-1-1=/datadisk1/salty.001.1
data-dir-prefix-1-2=/datadisk2/salty.001.2
data-dir-prefix-1-3=/datadisk1/salty.001.3
data-dir-prefix-1-4=/datadisk2/salty.001.4

4.Point 2 of the notes say:

• This SciDB cluster has a total of 8 instances, split evenly amongst the 8 cores and 4 disks.
But according to the configuration file, it seems there are 7+8=15 worker instances, if we don’t considering the other two servers( server-3 and server-4)

Can you help me understanding the example, thanks in advance.

Regards,


#2

You have caught a mistake from the guide: sorry about that. We’ve changed the example for 13.11, which is available now. The new example should be correct.

For point 2, you are correct: on server-0, there are 7 worker instances and 1 coordinator instance.

For point 3, you are correct: 8 worker instances. We meant to list 8 data-dir-prefix values, but mistakenly listed only 4 – again, this has been fixed for 13.11. Note, however, that you do not need to specify the data-dir-prefix parameter for each instance: you do not need to specify it all. If a server has multiple storage disks, and you want to assign more than one instance to each disk, then you must set the data-dir-prefix parameter for the instances on that server – otherwise, it is optional.

For point 4, your are correct, and the example was incorrect.