Hello all. After configuring the OWA/EV webapp we see the Enterprise Vault tab however when we click on it we get an "initialization failed" message. I read a few other community posts and some support threads and nothing has worked to correct the issue. EV is published externally as well but at this point we are just trying to get it working internally. Trying it internal and external has the same result error "initialization failed".
The environment is as follows:
x3 Exchange 2013 CAS/MBX servers SP1 RU8
2x EV v.11.01 (mailbox archiving, journaling)
F5 load balancer
Script we implemented:
$Mbx = get-mailbox "mailbox"
New-App -OrganizationApp -DefaultStateForUser:enabled -Url ("http://evarchive1.xyz.com/EnterpriseVault/OfficeMailAppManifest.aspx?LegacyMbxDn=" + $Mbx.LegacyExchangeDN + "&BaseURL=https://evarchive1.xyz.com/EnterpriseVault")
When running a trace in the browser i see the following error after clicking on the Enterprise Vault tab on a message in OWA
EVOMA (ERROR) 11:24:50.232: logAndReportFailure: EWSFindAssociatedItem.onEwsRequestComplete ID[ERROR_EWS_ERRORRESULTSTATUS] Err[Failed to send a request to the Exchange Server] [Result status:failed]
Things we have tried...
The Exchange Auth Cert thumprint matches an appears to be correct as per http://www.symantec.com/docs/TECH222757
We tried bypassing the Hardware Load balancer via host file and point direct to EV and Exchange as appropirate and get the same result.
We have tested EWS virtual direcotries on Exchange and have used other webapps such as bing maps and everything checks out ok. No errors.
When i run a dtrace i dont see anything - i believe its because i am not even making it to Exchange yet.
Also tried the basic things like restarting IIS on EV and Exchange servers, etc.
Any help would be appreciated.
Check for special characters, e.g. backslash or double quote, in the names and descriptions of your retention categories.
They could cause the "Initialization failed" error, and that issue is fixed in 11.0.1 CHF2.
We have a similar issue with EVOMA, its very useful for troubleshooting install Message Header Analizer App for test Virtual Directories in Exchange because thet apps are very similar.
I could ensure that app will not work. I suspect the problem is in the F5.
sadly, I have exactly the same issue - I already spent hours:-(
The setup is simple: W2012R2, EV 11.0.1 CHF2, 2x Exchange 2013 CU9, Kemp Loadbalancer.
We use different namespaces for internal and external access. First I tried just the internal function whithout bypass-URL, later on I tried with the external hostname -> same result.
What I did:
- bypassed loadbalancer with hosts files in order to work with just one Exchange server
- recreated Auth-certificate as described here: https://support.symantec.com/en_US/article.TECH222757.html
- besides that, Authconfig should be ok, as far as we use Lync without problems
- the URLs are correct, the certificates are all valid
- access to //evserver/EnterpriseVault//EnterpriseVault/OfficeMailApp.htm without errors (blank page)
- I also reviewed this post: http://www.symantec.com/connect/forums/failed-send-request-exchange-server-owa-2013 with no further clues
Error in Firefox (and IE):
logAndReportFailure: EWSFindAssociatedItem.onEwsRequestComplete ID[ERROR_EWS_ERRORRESULTSTATUS] Err[Failed to send a request to the Exchange Server] [Result status:failed]
initializationFailed(): Failed to send a request to the Exchange Server
What can I do?
Damn - I just changed DNS to point the external OWA URL to the internal IP (Kemp Loadmaster).
This worked so far!
The question now is: How we can accomplish the goal to use different URLs for internal and external use?
Besides this, I think about to use same external and internal URL for OWA only and use the TMG - I will give it a try.
You should ensure that the external URL is accessible from the intranet and have a valid SSL certificate on the EVserver for that url.
...I think I got it now (some pieces at least ;-)):
As mentioned, in one installation we work with different internal/external namespaces for all Exchange services (and so for OWA) -> "initialization failed" of the OWA mail app happens then.
In another setup, with split-DNS, we did not encouter this problem, everything works as expected from the beginning.
The funny thing is: I just configured the external hostname in the hosts-files on all Exchange servers to point to the internal IP address - this is enough, the app will be loaded and everything is fine now.
The more funny thing is: the Vault-server itself contacts the external Hostname as well by using the external hostname, not the internal. This happens at least for the processes:
EVExchangeWebServicesProxy, RetrievalTask and ArchiveTask
For that reason, I also configured the local hosts-file to use the internal IP for the Exchange servers just to optimise traffic between vault- and Exchange servers and not to re-route over the reverse proxy.
I really don't get it, but nevertheless, it works now.
Meanwhile MS suggests to use Split-DNS for Exchange, but sometimes you have to use different namespaces due to certificate issues (usage because of unregistered internal DNS-namespaces for AD) or just historical reasons.
It would be helpful to clear out the communication needs between the servers and services and why the external hostnames are preferred against the internal ones.
Just for information: I use certificates from an internal CA on all internal servers, the public one is bound just to the reverse proxy.
Thank you very much Buck,
I had the same issues and an Exchange Setup that is very close to yours. Adding the external namespace to Exchange Server's host file finally did the magic.
A shame for Verita's Support that let me wait for days without an answer. Using the internal URLs for internal communication should be the best solution for EV. But this job is up to Veritas.