Backups failing with Event ID 904
One of our clients was having an issue with backing up Exchange 2007 using Ahsay Online Backup Manager. The backups were failing nightly and Ahsay was returning the error CExBackup::backupService:HrESEBackupSetup: Error Number 0xc7fe1f45: Instance not found. A search of this error revealed that this error originated in Exchange and passed through to Ahsay. Checking the Application event log showed that for every failed backup there was an error with Event ID 904. The source is ESE Backup and the error text is Information Store (3396) Callback function call ErrESECBPrepareInstanceForBackup ended with error 0xC7FE1F45 Instance not found.
Researching this error online mostly pointed to this Microsoft Knowledge Base article, http://support.microsoft.com/kb/924204. However the article says this error occurs when the drive containing the transaction logs is full and I knew that this did not apply to the issue our client was having as they had nearly 245GB free on the drive where the transaction logs were stored.
I continued to search for answers and came across this posting from 2006; http://forums.msexchange.org/m_1800413118/mpage_1/key_/tm.htm#1800413118. In this forum thread the user Joseph Finnie reported that he observed this error when attempting to backup Exchange 2007 while a Recovery Storage Group was present on the server.
I checked the Exchange server for the presence of a RSG by opening the Exchange Management Console, selecting Toolbox and then selecting Database Recovery Management. This opens the Microsoft Exchange Troubleshooting Assistant. After the TA connects to Active Directory you will come to a page that says “Select one of the following tasks” and the section you want to look under is “Manage Recovery Storage Group.” If there is an RSG present, click the link for “Remove recovery storage group.” Once the RSG was removed this error stopped occurring and the backups completed successfully.
Registry editing through GPO
By Thomas Manier – Posted Feb 24, 2012 5:30PM
I wanted to give a quick tutorial on how to change registry keys through group policy. This is actually documented several times on the internet but very few give a clear understanding of the steps you need to take in order to achieve a quick solution. As a note, this will permanently add registry keys to computers. You can modify the values through group policy but it will require extra work in order to remove any key created in this manner. Also, I’m not going to cover everything you can do with this feature of group policy (technically, it’s an entire scripting language) but just how to push out registry keys to the HKEY_User directory.
VMDirectPath on Dell PowerEdge Servers with SD card for ESXi
We recently encountered an issue with VMware ESXi USB passthrough that is worth noting. On a Dell PowerEdge R and T series servers (T710, R710, R610, etc) you can opt to install ESXi on the embedded SD card. It works great and allows the OS disk to be completely separate from the datastore disk (which is great if problems arise). If you want to use VMDirectPath for USB passthrough then you need to be aware that one of the EHCI controllers is where the SD card is connected! The same controller that feeds the front 2 USB ports is also attached to the internal USB port on the motherboard and the SD card. If you pass that controller through then 1) the SD card becomes inaccessible to VMware and you will not be able to save any changes on the VMware host, 2) you will not be able to access the /bootbank directory, and 3) if you run “lsusb” you will not see any output.
Continue reading →
Music On Hold Streaming from an FXO port on CME 8.8
Here’s another CME feature that did not work quite as expected when reading the documentation. As with the conference bridge, I read along and followed the documentation from Cisco for “Configuring Music on Hold from a Live Feed” (the link is below) and the feature was still not working. When placed on hold, all callers heard was the default Cisco MOH audio file. This showed that the fall back portion of the config was working fine but overall something was missing.
“http://www.cisco.com/en/US/docs/voice_ip_comm/cucme/admin/configuration/guide/cmemoh.html#wp1010511
According to Cisco all that is required is a cable to connect to your live feed device, a fxo or e&m port, a dial-peer and an ephone-dn.
The cable we made on our own out of a bit of cat5, a RJ11 head and male to mail 3.5mm audio cable. Continue reading →
Configuring a conference bridge in Cisco CME
I recently ran into an issue configuring a conference bridge for a client. What makes this particular problem worth sharing is that I found the documentation provided by Cisco on this issue to be a little confusing and a little unhelpful.
The initial document I used to try to configure this issue can be found here: http://www.cisco.com/en/US/docs/voice_ip_comm/cucme/admin/configuration/guide/cmeconf.html#wp1021179.
In particular what confused me were the differences in step 5 of “Configuring SCCP for Cisco Unified CME” and step 3 of “Associating Cisco Unified CME with a DSP Farm Profile.” I have included the full configuration example of both of these sections below.
GoDaddy won’t issue a UCC cert for a Lync pool with a private domain name
A customer had a Lync 2010 environment with a pool name that used a private domain name and not an FQDN. When we attempted to request a UCC certificate from GoDaddy.com from the CSR generated from the Lync Deplyment Wizard we recieved the error:
You must use a fully-qualified primary domain name for UCC Certificate Request.
We called GoDaddy and according to them there was no way around this. So after attempting to generate a customized CSR with the Request-CsCertificate command without any success, we found how to manually generate a customized CSR using the certreq command.
Convert thick disk to thin on ESXi free
Here are the steps to convert a thick VMDK file to thin. Steps 2-4 are optional, but if you want to shrink it down as small as possible then don’t skip them!
- Evaluate how much space you have on the datastore. You will be creating 2 copies of the .vmdk file before we’re all done, so don’t fill up the datastore and crash all the VMs!
- Within the VM run a defrag of the disk. (this is optional but recommended)
- Download MS Windows Sysinternals ‘SDelete’ exe to the VM: http://technet.microsoft.com/en-us/sysinternals/bb897443
- Run ‘sdelete -c’ on the the disk. This will zero out unused space.
- Power offthe VM and make a note of what what datastore(s) the virtual disks are on and what they are named.
Continue reading →
Using Cisco 8945 IP phones with CME
We just finished testing the Cisco 8945 IP phone with CME using SCCP and it worked great. Point to point video works flawlessly. A few things to note in setting up these phones:
- Make sure you add the “video” command under telephony-service as well as under the ephone.
- We found that we could not call out our PRI with these phones once video was enabled. We had to add “bearer-cap Speech” under the voice-port to force speech only out the PRI. (The phones were trying to advertise video capabilities out the PRI and our service provider didn’t like that!)
- Although the phones are capable of sending and receiving 640×480 (VGA) video we were noticing that the max we were getting was 325×288 (CIF) resolution. We found that VT Advantage (which is what CME uses) maxes out at CIF resolution per this document.
Continue reading →
CME 8.8 upgrade and Exchange UM voicemail problem
We recently upgraded to the new Cisco CallManager Express (CME) version 8.8 so that we could try out the new 8945 phones for point to point video. Immediately after the upgrade Exchange no longer answered calls to voicemail or auto-attendants and instead gave us a busy signal. In reviewing the Exchange logs I found the following error:
Event ID: 1021
Source: MSExchange Unified Messaging
Task Category: UMCoreThe Unified Messaging server rejected an incoming call with the ID “2ACCC6EF-E2F911E0-9286A0ED-1857E3E@10.1.1.22“. Reason: “The Unified Messaging server received “INVITE” request with an invalid SIP header “USER-AGENT” with value “”.”
Centrally setup Windows XP machines to use the new 2008 Group Policy Preferences
With the new 2008 Group Policy Preferences there are some great new tweak tools to manage without the need for login scripts. Windows XP machines are not compatible with these new GPOs without first installing an update for Group Policy Preference Client Side Extensions (CSE). This update is downloadable from Microsoft as an executable via KB943729. Unfortunately Microsoft didn’t package this in a .MSI file so it can’t be deployed through the entire organization via a GP Software Distribution rule. If you have WSUS in place then it can be deployed there, but if not you’ll need to go back to good ol’ login/startup scripts to get this pushed out to all your machines. First some credit to a post over at serverfault.com by Evan Anderson who outlined the strategy. Continue reading →