hi, a nice author. I have a question that I want to make multi disk with individual fdbserver, but the fdbserver start unsuccessfully. I did this by following the steps below:
can you give me a detail config? Then, I find the ‘–dataFolder’ which type is ‘std::string’ in the fdbserver.actor.cpp, so I cannot understand how to config the ‘multi disk’ for big data
Multi-disk per machine is achieved by starting multiple fdbserver processes, one per disk. What is the error message? Look in journalctl -u foundationdb and also look for Severity="40" events in the xml files in /var/log/foundationdb/*.xml
Thank you. The error message in the fdbcli shows as follows:
127.0.0.1:4500 (reachable)
Unable to locate the data distributor worker.
Unable to locate the ratekeeper worker.
as you said ‘Multi-disk per machine is achieved by starting multiple fdbserver processes, one per disk’.
Question1:
if I configure single with five fdbserver with differ datadir in a machine as follows:
fdb> configure single
it means that multi disk config?
Question2: fdbserver1, fdbserver2, fdbserver3
if I configure double with three fdbserver with differ datadir in a machine as follows:
fdb> configure double
Which two fdbserver is the replicates fdbserver?
Does the third fdbserver store data? whate the role of the third fdbserver?
Q1: If you have a single machine, you can only have “single” replication mode (1 copy exists of the data).
Q2: If you have a single machine, and you configure 3 fdbserver processes, and select “configure double”, your cluster will fail, because you need at least 3 physical machines to enable that. If you did that earlier, I suggest you delete your test installation, data and reinstall (or add 2 additional machines to your cluster).