locked
Exchange 2010 Management Console.....Initialization failed RRS feed

  • Question

  • After a text book installation of exchange 2010, everything was working fine for about a day, then for no reason the management console gave the above error, with details... Connecting to remote server failed with the following message : The WinRM client received an HTTP error status (500), but the remote service did not include any other information.    The server is running 2008 R2 64.  I have tried allsorts to make it run, but to no avail, The mail system is working fine by the way ! all mail in and out is working with no errors....
    Thursday, April 1, 2010 5:27 PM

Answers

  • OK, so the mystery unravels itself !! Thanks Matt for your input, that filled in all the blanks that I had and while I didn't get the quick fix I wanted, i helped me realise why I had a problem.

    You are absolutely right on the second nic causing the problem, as thats exactly what I had, the only reason that I had such a problem with this is because the error reporting on exchange 2010 (for these and similar issues) is vague at best !  If the error had blamed the second nic, then I'm sure we would have got to the problem sooner. As for exchange getting 'confused' as stated by a microsoft engineer, this doesn't even suprise me the slightest. I think as yet real exchange 2010 'experts' are in short supply, give it another year and we may hear more proffesional answers.

    Going back to the original error message for other people reading this with similar problems, if your emc and ems won't connect giving the http error '500'. Don't look at one forum and think you have fixed it, there are many reasons for this error, see above posts, it is a generic error message and you need to look further (event logs etc) to get your real answer. I think the forums have covered most bases but this is the only one I have seen with this 'second nic' solution. 

    Bottom line ..don't fit a second nic onto a exchange 2010 server. It obviously raises issues that microsoft don't have the answer to. And in fact when you think about it, you should never have the need to do that anyway, if your reason for a second nic is to say; run sip traffic through...you shouldn't even be doing this on an Exchange server in the first place !

    I had to rebuild, which made me sad that there wasn't a true fix, however my advice when building exchange 2010, on server 2008 R2,  do it in stages, take backups at each stage so you can go back to a working stage....and only fit one network card !! (with IP V6 enabled)

    Tom

    • Marked as answer by Baronblue Thursday, April 29, 2010 7:51 AM
    Thursday, April 29, 2010 7:51 AM

All replies

  • There are a couple possible solutions in this post:

    Troubleshooting Exchange 2010 Management Tools startup issues
    http://msexchangeteam.com/archive/2010/02/04/453946.aspx

    Thursday, April 1, 2010 5:49 PM
  • I've checked out the thread, one of the symptoms noted describes exactly the error I am seeing, however when I checked out both the suggested possible causes, my setup did not have either of the lines missing. In addition he said that event 10113 was a symptom, and I don't get that..

    I have just noticed something else, OWA won't go online, I'm guessing that the two must be related, any more idea's ?

    Thursday, April 1, 2010 8:28 PM
  • Hi,

    "OWA won't go online"

    Do you mean there is an error message while you log in to OWA? Could you please post error message here?

    Did you install any software before the error appear?

    Please check whether the Powershell is point to "MsExchangePowerShellAppPool" in IIS Manager.

    Please check Server Manager->Feature->WinRM IIS Extension whether installed or not. If not installed, please install it. If already installed, please reinstalled it.

    By the way, have you reboot the server? It maybe works just by a reboot. 

     


    Frank Wang
    Monday, April 5, 2010 8:44 AM
  • Hi

    Sorry I was a little ambiguous there, the error I get for OWA is

    Outlook Web App didn't initialize. If the problem continues, please contact your helpdesk.
    The Microsoft Exchange Active Directory Topology service on server localhost can't be contacted via RPC. Error 0x5.

    I will re-instll WinRM, as it is already installed, tried re-boots, but no change, fortunately this server isn't live yet, so I have plenty of freedom at the moment.

    Also, there were no changes made to the server when the error first started, it was a fresh install, which was working fine for about 36 hours, then it just did this for no apparent reason

    Monday, April 5, 2010 11:22 AM
  • I have had the exact same issue for a month now. I have done 3 "bare metal" (OS and all) reinstalls, but it does the same thing to me. Works intially, though OWA doesn't work from "outside" even though the default IIS page does, then the EMC goes down and no amount of reboot and reinstall works. I can tell you a reinstall of WinRM won't help either. I have been trying to get Microsoft support to respond for two weeks, to no avail. Maybe you'll get an answer?
    Tuesday, April 6, 2010 3:09 PM
  • Glad to know I'm not alone, and from what I've been reading this looks like a common problem, I did try the WinRM re-install, and like you, I am on the third install from flat. I have access to a working system tonight and another engineer, we will be looking at this together, if I have any news I will post it here.

    Tuesday, April 6, 2010 7:09 PM
  • Yeah. it took three "bare metal" installs and 3 different LM sessions with M'soft engineers just to get it to complete an install from RTM media. It's taken another 3 "bare metal" installs trying to get Exchange to run, without success thus far. And as soon as the install (finally) completed Microsoft washed their hands of it and walked away, and now don't want to talk to me unless I either pay for their over-priced "software assurance" or give them a credit card number.

    Meanwhile, did you notice that by now you can't even connect from Outlook "inside". So it's not just OWA that "tanks" after a day or two - it appears the entire install (including power shell and emc) stops responding. This is the point I got to last time, when M'soft engineers would give no other response than to flatten/reload again, which I did. Like you, I left the installation alone over the balance of a weekend, and like you I discovered Monday morning that I couldn't proceed with configuration because the EMC was down. During that entire time the default IIS 7 page continues to display both internally and externally, so when they blamed it on IIS they were wrong as well. I have received no further guidance from them on what to try next, and so have filed a formal complaint for them closing the original case ID before the software actually ran as advertised. I hope you have more success with them than I did, but frankly I'm afraid you won't.

    Wednesday, April 7, 2010 10:55 AM
  • Anyone have any luck with this yet? Kind of tired of waiting 3 months for a working piece of software!

     

    Thanks!

    Matt

    Friday, April 9, 2010 11:25 AM
  • Matt...In answer to your previous post, my outlook is working fine, and is good and stable, with working email. While this is no use to me without administration, it gives me something to work on.

    If your whole exchange system fails after a short time, you may consider a hardware incompatibility, which initially works and them may fail after updates have been run. Without seeing more details I'm just guessing.

    Getting back to the exchange shell not connecting, I'm not really any closer to finding an answer, however, it would appear that this message can appear for different reasons, We haven't just found ours yet !!

    Sunday, April 11, 2010 10:17 PM
  • I agree, the Outlook "thing" is probably unrelated. Since I never had the EMC working long enough to properly configure the mail dbs, I'm not surprised frankly. It is definitely NOT a "compatibility issue". Nor have any of the reasons proposed by Microsoft for the consistent EMC/OWA failure thus far proved to be true/correct. I'm still waiting to hear from a couple of former consultants who have also attempted to install 2010 RTM, to see if they ever succeeded. They were the folks who first warned me about how bad this was, so I'm hoping maybe they found a fix on their own I can apply to the issues you and I are both seeing. I will of course pass anything they have to say on here for your use, and I hope you will do the same? I frankly don't expect much help from Microsoft on this, so I think we're on our own at this point. I am considering/discussing with those same consultants a contract to have them assist me, if they in fact have been successful in getting this working anywhere else yet. I will of course share any assistance they can provide here, as well. Thanks Matt
    Monday, April 12, 2010 11:40 AM
  • I had the same problem. I had toyed around with the IIS binding. Check it out and see if you can bind to all available IP. I had made some changes on that and that's what screwed it up... hope that helps...
    Thursday, April 15, 2010 8:30 PM
  • What I've been able to discover is that enabling ANY IP on a second NIC on the Exchange server is enough to crash the EMC after a day and a half or so. In my case IIS continued working all the way through , even when OWA and the EMC failed. In working with a Microsoft engineer, who seemed more interested in blaming everything else on my network and/or server than actually fixing the problem, I was told the following:

     

    Contrary to some postings on forums, IP V6 MUST be enabled, at least on the primary NIC, or Exchange can't talk to the domain and so will fail to start after a reboot, or after running a couple of days.

    If you enable IP V6 on any other NIC besides the primary, you will likely no longer be able to reboot your server in anything less than about 45 minutes. It's also probable that your Network Sharing Center console will no longer work once it does finish rebooting.

    According to the M'soft engineer, who I ended up hanging up on after I got fed up with listening to him blame my existing, fully functional for 2 years network for the ____ software bugs we were supposed to be working on, if you enable a second NIC, for say an outside connection to the server on your public Internet firewall, Exchange gets "confused" (in his words) and fails. My contention, and my request to the vendor who sold me this software and 150 CALs to go with it, is that this constitutes a defect in the product, makes it unfit for the purpose intended, and therefor entitles me to a full and immediate refund.

    That vendor, BTW, is the consultant I was trying to work with since M'soft won't provide support. They have no other site running this software in production, and I have yet to hear of one either. I'm thinking it's time to fall back to 2007, or maybe Scalix or a similar 3rd party package, and leave 2010 to M'soft to (hopefully) get fixed some day. Of course, I had the same hopes for Vista at one time, but that's a different story.

    Meanwhile, check your NIC configuration if you're still having these kinds of issues. If you don't mind disabling other NICs on your server to run this, it will probably be okay, though I haven't gotten it to run long enough to be sure of that yet. Don't take any advice about fiddling with IIS - it's so finicky that anything besides the settings input by Exchange during install will probably crash it, and you may very well have to reinstall the OS and Exchange to get it working again. I had one of the M'soft engineers do that to me once. Myself, I'm not willing to accept a half-baked product, especially with Exchange's price tag, and attempt to put it in production. Thanks for your help and advice

    Friday, April 23, 2010 11:24 AM
  • OK, so the mystery unravels itself !! Thanks Matt for your input, that filled in all the blanks that I had and while I didn't get the quick fix I wanted, i helped me realise why I had a problem.

    You are absolutely right on the second nic causing the problem, as thats exactly what I had, the only reason that I had such a problem with this is because the error reporting on exchange 2010 (for these and similar issues) is vague at best !  If the error had blamed the second nic, then I'm sure we would have got to the problem sooner. As for exchange getting 'confused' as stated by a microsoft engineer, this doesn't even suprise me the slightest. I think as yet real exchange 2010 'experts' are in short supply, give it another year and we may hear more proffesional answers.

    Going back to the original error message for other people reading this with similar problems, if your emc and ems won't connect giving the http error '500'. Don't look at one forum and think you have fixed it, there are many reasons for this error, see above posts, it is a generic error message and you need to look further (event logs etc) to get your real answer. I think the forums have covered most bases but this is the only one I have seen with this 'second nic' solution. 

    Bottom line ..don't fit a second nic onto a exchange 2010 server. It obviously raises issues that microsoft don't have the answer to. And in fact when you think about it, you should never have the need to do that anyway, if your reason for a second nic is to say; run sip traffic through...you shouldn't even be doing this on an Exchange server in the first place !

    I had to rebuild, which made me sad that there wasn't a true fix, however my advice when building exchange 2010, on server 2008 R2,  do it in stages, take backups at each stage so you can go back to a working stage....and only fit one network card !! (with IP V6 enabled)

    Tom

    • Marked as answer by Baronblue Thursday, April 29, 2010 7:51 AM
    Thursday, April 29, 2010 7:51 AM
  • It has long been standard practice to not multi-home Exchange servers.  The only two scenarios I can think of for a second NIC are (1) for use with network load balancing on a CAS and/or HT server, and (2) when a separate backup network is used, in which case you want to configure a private address on that network and static routes if necessary.
    --
    Ed Crowley MVP
    "There are seldom good technological solutions to behavioral problems."
    .
    "Baronblue" wrote in message news:e0e94443-bde1-49bf-9b9b-742df1e7a802...

    OK, so the mystery unravels itself !! Thanks Matt for your input, that filled in all the blanks that I had and while I didn't get the quick fix I wanted, i helped me realise why I had a problem.

    You are absolutely right on the second nic causing the problem, as thats exactly what I had, the only reason that I had such a problem with this is because the error reporting on exchange 2010 (for these and similar issues) is vague at best !  If the error had blamed the second nic, then I'm sure we would have got to the problem sooner. As for exchange getting 'confused' as stated by a microsoft engineer, this doesn't even suprise me the slightest. I think as yet real exchange 2010 'experts' are in short supply, give it another year and we may hear more proffesional answers.

    Going back to the original error message for other people reading this with similar problems, if your emc and ems won't connect giving the http error '500'. Don't look at one forum and think you have fixed it, there are many reasons for this error, see above posts, it is a generic error message and you need to look further (event logs etc) to get your real answer. I think the forums have covered most bases but this is the only one I have seen with this 'second nic' solution. 

    Bottom line ..don't fit a second nic onto a exchange 2010 server. It obviously raises issues that microsoft don't have the answer to. And in fact when you think about it, you should never have the need to do that anyway, if your reason for a second nic is to say; run sip traffic through...you shouldn't even be doing this on an Exchange server in the first place !

    I had to rebuild, which made me sad that there wasn't a true fix, however my advice when building exchange 2010, on server 2008 R2,  do it in stages, take backups at each stage so you can go back to a working stage....and only fit one network card !! (with IP V6 enabled)

    Tom


    Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."
    Thursday, April 29, 2010 3:21 PM
  •  

    My exchange 2010 sp1 server has two NICS.  We are a hosted exchange company and always assign a 2nd NIC on a separate subnet and separate switch for backup purposes ONLY.

     

    Restarting IIS fixed my problem, although it comes back time to time. I suspect this is a bug that Microsoft will fix soon.

     

    BTW, for those of you trying to run multiple organizations on 1 exchange server, exchange "tenant" mode now allows this and is really incredible.  It will save us a lot resources now that this feature is finally available.

     

    Tuesday, November 16, 2010 3:30 PM
  • That's not an unusual two-NIC configuration, but I presume you are using static routes to ensure that the backup NIC is used only for the specific subnets.  That's not really a multi-homed configuration that gives problems when it's properly configured.


    Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."
    Saturday, November 27, 2010 4:30 AM