powered by Jive Software

Contact List Sharing Issues

I recently setup an Openfire server and all of my clients are using Spark. I have 3 groups setup comprising of 17 users and I am using the internal database.

I love the Openfire and Spark products, but the contact list sharing tends to act very strange at times. If I change any of the sharing settings, add/remove users, create additional shared groups, etc it tends to reek havoc. Sometimes it will cause shared groups to totally dissapear from other group members lists, sometimes it will cause certain members of a group to disapear from the group (even though they are still a member) as well as other random issues. Sometimes the fix is to turn of sharing for a particular group then turn it back on, other times I have to remove and re-add members to a group, and sometimes I just keep messing around until it appears correctly. I’ve even had an issue where I have removed a user from all groups, but yet those groups still appear on their roster as “shared groups”.

Once I do get everything setup correctly everything is great and it is as solid as a rock, but as soon as I make a change I really have to fight with it to get it back to normal. Has anybody else experienced anything like this? I would love some advice/suggestions on getting it fixed as I really do like the Openfire/Spark products!

If a group is set to be shared to All, then it will appear in any user’s roster no matter belongs that user to any group or not. It means that group is shared to every user on a server.

How did you determine that groups are messed up? Did you look in Spark or do they look messed up in Admin Console > Groups too? If it is only in Spark, then maybe it is sufficient to restart Spark. Or maybe you can try restarting Openfire after you have made all changes. Sometimes i see something similar in Spark, when i’m doing changes in rosters, though maybe not so drastic as you have described, but usually client’s restart solves this. Spark just can’t update changes for some reason, maybe because Openfire is not pushing changes right away nad keeping them in caches. That’s why server restart should help i think. It will write every change into database and then Spark should pull normal rosters.

I determined the groups were messed up by looking within Spark, but I did not look at the admin console to see what it said. I did try simply restarting the spark client, but that didn’t resolve the issue. I have not tried restarting Openfire however. The next time this happens I will try a client restart as well as a server restart and see if either of those will help. Thanks for the help!

I have some additional info to add on this subject.

I had three groups previsouly setup. Lets call the Group A, B, and C. I created a brand new group called D. As I was adding users to group D they were showing up in group C! I checked the user roster from the admin pannel and they were showing up as members of group C even though they weren’t at all. They were also showing up in Spark as being members of group C when they weren’t at all. I restart of Spark didn’t resolve this issue although a restart of Openfire did. Once Openfire was restarted everything appeared just as intended with the users in their correct groups. I tried adding more users to group D and experienced the exact problem with a restart of Openfire fixing the issue.

Seems like a very odd problem, but I happy to say at this point restarting Openfire solves the issue.

Well, i don’t know what could be causing this. I’m using Openfire with embedded-db and usually it is not restarted for months and i sometimes move some users from group to group (usually i add them first in other group and then delete them in the previous). Usually it takes only Spark restart to update that info and i have never seen users dropping into wrong group in the Admin Console. Though, i haven’t created any new group for a while. As i said this could be an issue with Openfire caching changes, and then it probably gets messed up then trying to pull information about the rosters, because some infomation is only in caches and some info is in a database already. Maybe when you add a new group, it holds it in cache, but when you add users to that group it checks with a DB and can’t find that group in there and then puts them in old group. Indeed weird issue and nothing we (users) can do about it. At least there is a workaround with a server’s restart.

You can do an experiment. Create a new group and wait for a few hours, then try adding new members to it. Check how it behaves. Maybe this group will be written to a database after some time and then maybe there won’t be such problems with roster, at least in Admin Console.