#49559 lib389 topology_m2 does not always assign the same replicaids to instances
Closed: wontfix 2 years ago by vashirov. Opened 4 years ago by lkrispen.

when running a testcase with topology_m2 I noticed that master1 has replicaid: 2 and master2 has replicaid: 1.
This is confusing but could be handled. But later I noticed that this assignment can vary, in on test run I do get:

 ldapsearch ....-b "cn=config" objectclass=nsds5replica nsDS5ReplicaId
 dn: cn=replica,cn=dc\3Dexample\2Cdc\3Dcom,cn=mapping tree,cn=config
 nsDS5ReplicaId: 2

and in the next:

l ldapsearch ....-b "cn=config" objectclass=nsds5replica nsDS5ReplicaId
dn: cn=replica,cn=dc\3Dexample\2Cdc\3Dcom,cn=mapping tree,cn=config
nsDS5ReplicaId: 2

so when trying to find out the differences between a good and bad run, starting from csns I was always looking at the wrong logs.
Tests need to be identical for each run, and it would be good if replicaids would be aligned with server ids. I know for replication itself this doesn't matter, but for debugging it is good to have.


In the replication suite update the way replica id's was assigned changed to the order in which the replicas are created rather than being statically assigned. This is a change to simplify the external interface for admins so that replica ids are assigned that are always unique and valid, and prevents duplicate rid assignment which would create an invalid ruv.

As a result the first instance created will get replica id 1, the second id 2.

This indicates a race condition or other issue in the stability of the test if these are not being allocated "in order" as you expect. What test case triggered the issue for you so that I can investigate when I return from PTO?

Metadata Update from @firstyear:
- Custom field component adjusted to None
- Custom field origin adjusted to None
- Custom field reviewstatus adjusted to None
- Custom field type adjusted to None
- Custom field version adjusted to None

4 years ago

I reported this for the latest test Simon was creating for #49043.

But I have seen this with the basic replication acceptance suite with 4 masters, where the rids did not match the server ID

Metadata Update from @mreynolds:
- Issue set to the milestone: 1.4.0

4 years ago

Metadata Update from @vashirov:
- Issue close_status updated to: fixed
- Issue status updated to: Closed (was: Open)

2 years ago

Metadata Update from @lkrispen:
- Issue close_status updated to: worksforme (was: fixed)

2 years ago

389-ds-base is moving from Pagure to Github. This means that new issues and pull requests
will be accepted only in 389-ds-base's github repository.

This issue has been cloned to Github and is available here:
- https://github.com/389ds/389-ds-base/issues/2618

If you want to receive further updates on the issue, please navigate to the github issue
and click on subscribe button.

Thank you for understanding. We apologize for all inconvenience.

Metadata Update from @spichugi:
- Issue close_status updated to: wontfix (was: worksforme)

2 years ago

Login to comment on this ticket.

Metadata