#467 CLEANALLRUV abort task should be able to ignore down replicas
Closed: Fixed None Opened 7 years ago by mreynolds.

Current behavior is that the abort task will run forever until all the replicas have received the task, but there are times when you might want to abort because a replica is down. Need to add a new option to the task to handle such scenarios.


Also discovered a crash when running the abort task against a server that doesn't have replication setup. This is fixed in this patch.

Looks good, except for this:
{{{
2370 if(strcasecmp(data->certify,"yes") == 0){
}}}
It looks as though data->certify can be NULL here, so you probably want to check for that.

Replying to [comment:2 rmeggins]:

Looks good, except for this:
{{{
2370 if(strcasecmp(data->certify,"yes") == 0){
}}}
It looks as though data->certify can be NULL here, so you probably want to check for that.

Yes, there was one case where it could be null/undefined, and that's when the server would of been shutdown in the middle of the task, and resumed at startup. So this was fixed in repl5_replica.c.

git merge ticket467
Updating 1d6dd39..ce40822
Fast-forward
ldap/servers/plugins/replication/repl5.h | 1 +
ldap/servers/plugins/replication/repl5_replica.c | 4 +
.../plugins/replication/repl5_replica_config.c | 68 ++++++++++++++-----
ldap/servers/plugins/replication/repl_extop.c | 5 ++

git push origin master
Counting objects: 19, done.
Delta compression using up to 4 threads.
Compressing objects: 100% (10/10), done.
Writing objects: 100% (10/10), 2.16 KiB, done.
Total 10 (delta 8), reused 0 (delta 0)
To ssh://git.fedorahosted.org/git/389/ds.git
1d6dd39..ce40822 master -> master

Metadata Update from @mreynolds:
- Issue assigned to mreynolds
- Issue set to the milestone: 1.2.11.15

2 years ago

Login to comment on this ticket.

Metadata