Monitoring Exchange 2007 SCR with Ops Mgr 2007
I was a bit surprised to see that the most recent Management Pack for Exchange Server 2007 for Operations Manager 2007 didn’t understand the differences between CCR and SCR…and how those differences would impact the log queue depths for replication. So, if you’re going to implement SCR (or you already did), be sure to override the thresholds for the following monitors for your SCR targets:
- MSExchange Replication: ReplayQueueLength - sustained for 5 minutes - Red(>15)
- MSExchange Replication: ReplayQueueLength - sustained for 5 minutes - Yellow(>7).
Since the replay queue length is always going to be rather large due to the nature of how SCR operates (this value will never be less than 50 logs and will almost always be more than that), you’ll need to determine what the normal values of the replay queue depth are for your SCR targets. As an example, I have two SCR targets (from CCR sources) that routinely have replay queue depths of 300 - 400 logs with just a 1 hour ReplayLagTime value configured! See this TechNet article for more information configuring SCR.
I’d suggest making your “yellow” value at least 100 above your normal average replay queue depth and your “red” value at least 250 - 350 above your normal average. Make your overrides on the specific mailbox servers that are your SCR targets…not at the global level.
Data Protection Manager 2007 Management Pack updated
Version 2.0 of the Data Protection Manager 2007 management pack for Operations Manager 2007 was recently released.
This is is not a converted version, so installation is easy. Be sure to remove the previous version if you had it installed.
On the DPM side of things, make sure that you have SP1 or the Feature Pack in KB949779 installed before importing the management pack. As usual, get your management packs in the Catalog.
This management pack offers end-to-end monitoring of both the DPM servers and clients.
Ops Mgr DNS Management Pack updated
The DNS Management Pack for Operations Manager 2007 has been updated again to version 6.0.6480.0, but it still doesn’t seem to fix the the major complaint with the previous version (6.0.6278.27).
The DNS Library WINS TTL Monitor, despite getting some updated verbiage in the MP Guide, still appears to be broken. The guidance now specifically mentions that you need to create an A record in DNS, and provides some override information. No matter though, even with creating the A record “placeholder” in the forward lookup zone that corresponds to the domain name of the Ops Mgr management server (i.e. placeholder.mydomain.local), the alert still comes in. Refer back to my previous post about that item. So…this monitor gets overridden again.
Ops Mgr “The WINS Connector has stopped working”
Seems that the v6.0.6278.27 DNS management pack for Operations Manager 2007 has its own bug of fame. The “The WINS Connector has stopped working” alert will likely become familiar to you courtesy of the DNS Library WINS TTL Monitor. It seems that the MP ships with the “hostname” field filled in with “Placeholder”…but if you read the relase notes you’ll know this and know to fill it in with the NETBIOS name of something on your network or to create a WINS record for the host named PLACEHOLDER.
Setting the WINS TTL Monitor
The Windows Internet Name Service time to live (WINS TTL) monitor performs an NSLOOKUP query for a test (A) record named “placeholder” to verify that WINS resolution is working properly. You will need to create a test record called “placeholder” or create an override to change the target name.
Setting the Global Zone Resolution Monitor Target
The global zone resolution monitor performs an NSLOOKUP query for a test (A) record named “placeholder” to verify that global zone resolution is working properly. You will need to create a test record called “placeholder” or create an override to change the target name.
However, what really seems to happen is no matter what you do (or at least no matter what I did), you’ll still get the “The WINS Connector has stopped working” alerts. I tried individual overrrides, I tried a type override, I used upper case hostnames, lowercase hostnames, IP addresses, even tried creating a WINS record named PLACEHOLDER. No joy.
My solution: disable the monitor (DNS Library WINS TTL Monitor) until Microsoft updates the management pack. Have you been able to get it working?
Ops Mgr 2007 “Failed Accessing Windows Event Log”
If you’ve been getting these errors from your Operations Manager 2007 DHCP Management Pack, the solution is at hand! Here’s the typical error you’d get:
Failed Accessing Windows Event Log
The Windows Event Log Provider was unable to open the System event log on computer ‘{69E3F0CF-36C3-9FDC-1892-42FB003E34DB}’ for reading. The provider will retry opening the log every 30 seconds. Most recent error details: The RPC server is unavailable. One or more workflows were affected by this. Workflow name: Microsoft.Windows.DHCPServer.Library.Server.UnitMonitor.DependentServiceHealth Instance name: myserver.mycompany.local Instance ID: {69E3F0CF-36C3-9FDC-1892-42FB003E34DB} Management group: MyCompany-SysCtrOpsMgr
It seems that the DHCP MP v6.0.6383.0 introduced this annoying bug that seemingly had no solution. Well, I guess it really didn’t have a solution…until the next version was released. That version, v6.0.6452.0, was released in November and seems to fix the problem. Before you import this updated version though, be sure to remove any exisitng DHCP management packs. Per the included documentation:
Before you import the Windows Server DHCP Management Pack, take the following actions:
- Perform a full backup of the Operations Manager 2007 database.
- Uninstall any existing DHCP management packs. You may wish to record custom overrides and rules for possible inclusion in the new DHCP management pack.
Of course, the fix for this bug isn’t documented in the release notes, but none the less it does appear to be fixed.
Exchange and SQL OpsMgr Management Packs Updated
Updated Management Packs for Exchange Server 2007 and SQL Server were recently released for Operations Manager 2007. Check out the updates below as listed in the Management Pack documentation. As usual, download your Management Packs from the MP Catalog.
The following list contains the issues that are fixed in version 6.0.6461.0 of the Exchange Server 2007 Management Pack for System Center Operations Manager 2007:
- Updated the display name of the top level folder in the Operations console to “Exchange Server 2007 (Converted)”.
- Changed how the management pack monitors disks. Previously, disk monitoring rules lacked alert suppression. The following rules and monitors were disabled:
- Disk % Free Space low - Red(<10%)
- Disk % Free Space low - Yellow(<15%)
- Disk Free Megabytes low - Red(<20MB)
- Disk Free Megabytes low - Yellow(<40MB)
The following rules were added:
- Disk Free Space Low Red Rule (by default alerts if any disk has less than 20MB free space)
- Disk Free Space Low Yellow Rule (by default alerts if any disk has less than 40MB free space)
- Disk Percentage Free Space Low Red Rule (by default alerts if any disk has less than 10% free space)
- Disk Percentage Free Space Low Yellow Rule (by default alerts if any disk has less than 15% free disk space)
The new disk monitoring rules support alert suppression. It is also possible to use the Server Operating System Management Pack to monitor Exchange disk space, if desired. In that case, the above rules can be disabled.
SQL Server Management Pack version 6.0.6460.0 includes the following changes:
- The DB discovery script no longer casts the values that correspond to “Database Size (MB) (Numeric)” and “Log Size (MB) (Numeric)” as INT, to avoid overflows exceptions within the script itself.
- The DB discovery script now checks for possible overflow on the “Database Size (MB) (Numeric)” and “Log Size (MB) (Numeric)” values and prevents those overflows from occurring.
- For the numeric properties on the database class we are limited to 2147483647 MB (~2047 TB). In the event that a DB or log file exceeds that size then the value will be set to the maximum possible value of 2147483647 to prevent overflows. In these instances the ”Database Size (MB) (String)” and “Log Size (MB) (String)” will support larger values.
Updated Ops Mgr 2007 SQL and Windows Management Packs
A couple of recently updated Management Packs for Operations Manager 2007…SQL and Windows Server Operating System. Here’s the updates from the included documentation in the Management Packs. As always, get your Ops Mgr Management Packs here.
The October 2008 release (version 6.0.6321.5) of the Windows Server Operating System
Management Pack includes the following changes:
- Various scripts were updated to prevent issues with locale specific number formatting.
SQL Server Management Pack version 6.0.6441.0 includes the following changes:
- General Changes
- The management pack now includes the SQL Server 2008 discovery and monitoring management packs. SQL Server 2008 monitoring is identical to the SQL Server 2005 management pack, including the new functionality added in this release for SQL Server 2005 monitoring.
- Addressed issues with a few performance rules attempting to collect performance counters or instances by the wrong name.
- Updated a number of discoveries, rules, and tasks to ensure they use the correct RunAs profiles for discovery and monitoring.
- Updated the criteria on a number of event-based rules to make their criteria more specific to reduce alert volumes.
- SQL DB Engine discovery will now work on systems that do not have the SQL tools installed.
- The SQL Server 2005 and SQL Server 2008 management packs support discovery and partial monitoring of 32-bit instances of SQL components installed on 64-bit operating systems. This is discussed in more detail in the “Supported Configurations” section of this guide.
- All monitors are now set to public accessibility, allowing increased customization. Most notably, custom diagnostics and recoveries can now be added.
- The “Database Size” and “Log Size” properties of databases are now available as a numeric property, in addition to the string form of the property that already existed.
- Updated knowledge for various reports.
- Database snapshots are no longer discovered and monitored as if they were a full-fledged database.
- The possible states of the “Blocking SPIDs” monitor in all management packs have been changed to better align with the default override behaviors. Previously, the monitor could either be in a “Success” or “Warning” state. The “Warning” state has been replaced with “Error”.
- Changes to the SQL Server 2000 Management Pack
- The default thresholds for the “Job Duration” have been revised. The numbers were previously an integer corresponding to minutes when it should have been in the format of HHMMSS. The knowledge for this monitor has been updated to more clearly explain the format of the threshold.
- Changed the frequency at which the SAPasswordMonitor.vbs script is run from 24 seconds to 24 hours.
- Changes to SQL Server 2005 Management Pack
- Added monitoring for log shipping with the “Destination Log Shipping Monitor” and “Source Log Shipping Monitor” monitors.
- Fixed a few typos and formatting issues with product knowledge in various places.
Updated Windows Server and Terminal Services Ops Mgr MPs
New Management Packs for Windows Servers (2000, 2003 and 2008) and Terminal Services (2000, 2003 and 2008) were also released in late July.
Update your Windows Server MP first, get it here, which is version 6.0.6278.22.
What’s New
The following features are new in this release of the Windows Server Operating System Management Pack:
- Introduced support for Windows Server 2008.
- Added version specific notations to the names of the performance collection rules to reduce ambiguity, specifically when search for rules by name.
- Added a new task to leverage the new “Admin” switch of the remote desktop connection tool. More detail is provided in the “Troubleshooting” section.
Changes in This Update
The update to the June 2008 version of the Windows Server Operating System Management Pack includes the following changes:
- Updated the logical drive discoveries to omit mapped network drives.
- Addressed an issue with the Logical Disk Free Space to prevent it from looking like the thresholds were set incorrectly.
- The state of the configuration of a server now reflects the state of its operating system as well.
- Fixed an issue with the Server Service Configuration Health monitor which prevented it from ever generating an alert.
You can get the Terminal Services updated MP here, which is also version 6.0.6278.22. No “fixes” listed in this MP, but it does add support for 2008 based Terminal Servers.
What’s New
The following features are new in this release of the Terminal Services Management Pack specific to Terminal Services in Windows Server 2008:
- TS Gateway
- TS Session Broker
- TS Web Access
Exchange 2007 Ops Mgr MP Updated
Version 6.0.6363.0 of the Exchange 2007 MP was released in late July (with no fanfare that I noticed).
You can download it here.
Oddly though, the only fix reported in the update is listed as being for Exchange Server 2003?
Fixed an issue where clustered Exchange 2003 mailbox servers were discovered as being of type Ex.Common. Please see the Known Issues section for more information on how to resolve this issue.
Terminal Services Self-Tuning Monitors, Part 2
As a follow-up to my earlier entry about tuning STT monitors, specifically for Terminal Services…
Either our Terminal Services environment is really, really, really out of the normal (a little bit due to our applications in use, but not excessively so…) or the STT monitor for TS sessions (active and inactive) just plain doesn’t work.
Bottom line, I’ve disabled both monitors after attempting to find an override point for three weeks in a row that would give valid data. No matter what set of thresholds are chosen, like clockwork, at the end of 7 days all of the terminal servers start to issue alerts. I’m marking this one as a bug and hoping for an updated MP in the near future for Terminal Services.
Anyone else have similar experiences?
-
Translator
