How to create a DR site?

Using FoundationDB
1

Hello!

I’m trying to create and and add a DR site for my fdb cluster. Unfortunally, there is no any step-by-step primer in the ‘Administration’ section of the FDB Documentation, so I made some steps by own.

I made:

  1. Created a second three-node cluster. Checked with fdbcle that it is reachable
  2. Copied cluster files from both source and DR clusters to a client machine as fdb.cluster and standby-fdb.cluster
  3. Ran the command on the client machine (with fdb client installed):

fdbdr start -d standby-fdb.cluster -s fdb.cluster

And the response is:
The DR on tag `default’ was successfully submitted but no DR agents are responding.

  1. fdbdr status -d standby-fdb.cluster -s fdb.cluster

displays:

The DR on tag `default’ is NOT a complete copy of the primary database (just started).

  1. fdbcli -C fdb.cluster
    status details

shows
Backup and DR:
Running backups - 0
Running DRs - 0

Seems the data tansfering to DR is not in progress. What have I do more for copying data to the DR site?

2

You may want to try starting DR agents in the other direction as well, with standby-fdb.cluster as the source and fdb.cluster as the destination.

Edit: It’s unlikely to be what’s happening here, but it’s generally good to have them in both directions to help with switching and failover.

Edit: Oh, after re-reading it I see you’re not running DR agents in either direction. That’s something that we should make clearer in the documentation. You need to run the dr_agent binary, with the same source and destination parameters, and leave those running as background processes. The dr_agent process is responsible for doing the work of copying the data.

3

You are right. There weren’t dr_agents running. After starting dr_agents, the replication took a progress.

Makin a documentation clearer would be great.

4

Excuse me, sir, can you please explain how did you start the dr_agent? I have the same issues as you, I really appreciate it.

5

Now I start them by adding additional sections to /etc/foundationdb/foundationdb.conf

6

yes, I found this way too. now I can run dr_agent, but I face this error when I try to use fdbdr command:

SIGNAL: Segmentation fault (11)
Trace: addr2line -e fdbdr.debug -p -C -f -i 0xa9a655 0x7f3b39c89b20 0x65835b 0x65c214 0x4f9998 0x5eeaef 0x917698 0x84 94c8 0x84976f 0x846868 0x846b3d 0x84b4e0 0x8c22b8 0x87b7fc 0x87b90f 0x868ac8 0x866a18 0x866bcd 0x880d58 0x7d3228 0x7e c27b 0x9fd40b 0x9fd505 0x4f9900 0xaf5840 0x841712 0x4b3573 0x7f3b3934b493
Segmentation fault (core dumped)

did you face this by any chance, sir? thank you I really appreciate it.

7

I think you need to put the cluster file in the fdb_dr command in order to check the status, I think there was an issue logged about this.