Sign up for Office 365
Learn more about Office 365
I've succesfully setup a hybrid deployment between Exchange 2010 and Exchange Online for one of my customers.
All works well:
- mailflow is ok both ways
- addressbooks are ok in both environments
- shared calendars can be opened both ways
However, when opening a shared calendar in either environment, the updates of those calendars are happening very slow and on an irregular basis.
Does anyone know how this process goes? I can't figure it out.
When I put an appointment in an on-premise calendar, it can take anywhere from 5 minutes to over half an hour before I see that appointment in a cloud user's mailbox who has this shared on-premise calendar opened. This goes for both Outlook (2007 and 2010) and OWA.
Anyone any ideas on this?
Thanks for the feedback.
Thanks for your post. I fully understand your situation. As for this case, it's a preformance issue. For Office365 ,to query free/busy information cross organization is using EWS service. To isolate this case between local and Office365, I'd like you can enable hybird IIS log to check the timing of query request and response.
If it's not a case on local, please feel free to post it here.
Thanks, Neo Zhu
I'm seeing lines like this in logging (I replaced internal ip's and names with fake ones):
2012-05-23 12:52:37 192.x.x.x POST /EWS/Exchange.asmx Requester=S-1-5-21-2503399708-2711579577-3781570003-5668;federatedxforest=1;Threads.Worker.Available=198;Threads.Worker.InUse=0;Threads.IO.Available=200;Threads.IO.InUse=1;Failures=0;MailboxRPC.TimeTaken=1;MailboxRPC.RequestCount=12;AD.TimeTaken=0;AD.RequestCount=2;Request.CPU.Main=31;FederatedTokenCreationLatency=1462;LongPoleAutoDiscoverRequest.TimeTaken=407;LongPoleAutoDiscoverRequest.RequestCount=1;LongPoleAutoDiscoverRequest.Destination=pod51015.outlook.com/.../WSSecurity;Request.Phase.PreQuery=15;Request.Phase.RequestDispatcher.BeginInvoke=2781;Request.Phase.RequestDispatcher.Complete=859;Request.Phase.PostQuery=0;Request.CPU.Total=62; 443 domain\user 192.x.x.x Microsoft+Office/14.0+(Windows+NT+6.1;+Microsoft+Outlook+14.0.6112;+Pro) 200 0 0 3687
2012-05-23 12:57:30 192.x.x.x POST /EWS/Exchange.asmx Requester=S-1-5-21-2503399708-2711579577-3781570003-5668;federatedxforest=1;Threads.Worker.Available=199;Threads.Worker.InUse=0;Threads.IO.Available=200;Threads.IO.InUse=1;Failures=0;MailboxRPC.TimeTaken=0;MailboxRPC.RequestCount=12;AD.TimeTaken=0;AD.RequestCount=2;Request.CPU.Main=31;FederatedTokenCreationLatency=1529;LongPoleAutoDiscoverRequest.TimeTaken=129;LongPoleAutoDiscoverRequest.RequestCount=1;LongPoleAutoDiscoverRequest.Destination=pod51015.outlook.com/.../WSSecurity;Request.Phase.PreQuery=15;Request.Phase.RequestDispatcher.BeginInvoke=2984;Request.Phase.RequestDispatcher.Complete=1281;Request.Phase.PostQuery=0;Request.CPU.Total=31; 443 domain\user 192.x.x.x Microsoft+Office/14.0+(Windows+NT+6.1;+Microsoft+Outlook+14.0.6112;+Pro) 200 0 0 4281
Do these seem normal to you?
This is logging from the EWS site within IIS on the exchange server.
Yes, the speed is pretty good. I'd like to continue narrow down this issue. Based on the description of the log, I'd like you can logon to Office365 OWA and use scheduling assistant to query free/busy information of the local user. If the updated information can be listed immediately.
How are you doing? I am just writing to see if there is any progress on this issue. I do not mean to push you. I just want to ensure that you do not encounter any problems during the troubleshooting. Thank you.
My apologies for the delay, I was busy doing the migration.
The problem persists, so I tested your suggestion with OWA and the scheduling assistant.
I've put an appointment in the cloud based mailbox and then scheduled a meeting with that person. I can add the person without any problem but it doesn't show the updated calendar, I don't see the appointment I had put in.
apparently I overlooked some things.
So: using the scheduling assistant I immediately get the correct free/busy information. However, when I then return to the shared calendar, the appointment stays on the old time in stead of the new one. This both for OWA and Outlook.
So it seems as if the calendar itself doesn't update free busy?
Or am I incorrect?
Hi Bart Malfait - TENit,
Thanks for your reply and updated information, If scheduling assistant can immediately get the free/busy information, it means there is no problem for the hybird server and configuration.
For this case, scheduling assistant is using the different way with open another calendar. For opening another calendar is also related to the local network preformance.
So what are you saying now exactly? That the problem is network related? Why do I have the same issue on OWA then?
This doesn't make any sense?
I really need a solution for this issue.
2.Verify that your on-premises organization firewall isn't blocking connections to your coexistence server. The on-premises organization firewall should allow HTTPS connections (port 443) to the hybrid server.
3.Use the Test-FederationTrust –UserIdentity <cloud-based user mailbox> -Verbose command in the Exchange Management Shell on the on-premises hybrid server to verify that a delegation token is properly received for a cloud-based user.
4.Verify that your on-premises users have downloaded and configured their Office applications by installing the Microsoft Online Services Sign-in Assistant and updates for Outlook 2007 or Outlook 2010 email clients. This can be downloaded by logging into your Office 365 site and selecting the Downloads link in the Resource section of the home page. For your reference:
Below my answers:
1) is ok
2) is ok
3) is OK, below is the output:
RunspaceId : 76c9c4a6-20a3-4482-981c-822c2aa27f4e
Id : FederationTrustConfiguration
Type : Success
Message : FederationTrust object in ActiveDirectory is valid.
Id : FederationMetadata
Message : The federation trust contains the same certificates published by the security token service in its federat
Id : StsCertificate
Message : Valid certificate referenced by property TokenIssuerCertificate in the FederationTrust object.
Id : StsPreviousCertificate
Message : Valid certificate referenced by property TokenIssuerPrevCertificate in the FederationTrust object.
Id : OrganizationCertificate
Message : Valid certificate referenced by property OrgPrivCertificate in the FederationTrust object.
Id : TokenRequest
Message : Request for delegation token succeeded.
Id : TokenValidation
Message : Requested delegation token is valid.
VERBOSE: [14:41:24.645 GMT] Test-FederationTrust : Admin Audit Log: Entered Handler:OnComplete.
VERBOSE: [14:41:24.645 GMT] Test-FederationTrust : Ending processing &
4) done, no change
Additional information that could be usefull in troubleshooting:
1) free/busy is working both ways (as verified with the scheduling assistant)
2) when 'share calendar' from outlook 2007 on-premise is done, we receive a notification that the cloud user is not part of the organisation, hence the sharing invitation cannot be sent. When the same on-premise user does exactly the same from OWA 2010, the sharing invitation can be sent without issues. the cloud user can then, from OWA 2010 open up the shared calendar from the on-premise user. after this was succesfull, the calendar is also added in the outlook 2007 of the cloud user, and he can then also see the shared calendar from outlook.
3) when the on-premise user changes an appointment, it took up to 45min today to have this change reflect in the shared calendar in the cloud mailbox!!!
We VERY URGENTLY need a solution for this problem.
Hi Bart Malfait,
Thanks for your reply. For hybird deployment, we focus on available service to query the free/busy information. The way is different from open other calendar. As for this case, you can take a look the private message on the right panel.
I've created a support request, more info in your private messages.
Thanks , I have recived that. I will also update the information to SE who working on this case.
Have you found a solution to your problem?