DAG maintenance

All my DAG members are connected to a main switch which will undergo maintenance. Obviously all servers will lose network access so is it better to have all servers powered off/ services shut down during this period? I don’t want servers trying to fail
over databases constantly and generate alerts. I’m aware of the StartDagServerMaintenance.ps1 script but am I right in saying this should only be used when you actually want to fail over databases to another server? If all servers will be unavailable (planned
and agreed outage), does anything need to be done on the servers for preparation?

You can shutdown servers if you have an agreed window.
But do consider options for any incoming emails. Will your anti-spam solution queue the emails for delivery when the server is back online?

Similar Messages

Shutting down secondary DAG node

Two Exchange servers setup in a DAG.
I need to expand the log VM on our secondary Exchange server (2013). I am very new to DAG Exchange administration. I am aware there are scripts to initiate DAG maintenance. I am just curious if this can be
safely accomplished by going through the steps below?
Suspend each mailbox db through EAC on the primary Exchange server
Shutdown the secondary
Expand the log VM
Boot up the secondary Exchange server
Resume DAG replication through Primary EAC
Thank you,
-TSGzz

Thank you for the feedback. I found another good overview of what I will go through tomorrow morning. There are also some scripts that were recommended in the 2nd link.
http://exchangeserverpro.com/exchange-2013-installing-cumulative-updates/
http://vanhybrid.com/2013/11/28/script-putting-exchange-server-2013-into-maintenance-mode/
MANUAL PROCESS
Set Exchange server into DAG maintenance mode
Suspend the DAG member from the cluster
Suspend-ClusterNode –Name Ex13DR
Disable database copy activation.
Set-MailboxServer Ex13DR –DatabaseCopyActivationDisabledAndMoveNow $true
Review the existing database copy auto activation policy, so that you can return it to the same configuration after you’ve completed the upgrade.
Get-MailboxServer EX13DR| Select DatabaseCopyAutoActivationPolicy
DatabaseCopyAutoActivationPolicy : Unrestricted
Set the auto activation policy to “Blocked”. If the policy is already set to “Blocked” then there is no action required.
Set-MailboxServer EX13DR –DatabaseCopyAutoActivationPolicy Blocked
Put the server into maintenance mode.
Set-ServerComponentState EX13DR –Component ServerWideOffline –State InActive –Requester Maintenance
Taking Exchange Server Out of Maintenance Mode
Set-ServerComponentState EX13DR –Component ServerWideOffline –State Active –Requester Maintenance
Resume-ClusterNode –Name EX13DR
Set-MailboxServer EX13DR –DatabaseCopyAutoActivationPolicy Unrestricted
Set-MailboxServer EX13DR –DatabaseCopyActivationDisabledAndMoveNow $false
Set-ServerComponentState EX13DR –Component HubTransport –State Active –Requester Maintenance

How to upgrade Exchange Server 2013 CU5 on my site?

My site have four Exchange servers . The four Exchange Servers are installed mailbox and cas roles. The
Dag achieved by the four servers. Please help me to upgrade exchange server 2013 cu5 on the four servers. how to do? Thanks.
北京老马

Hi Martin
Find the below information.
Upgrade Sequence for each Role:
==========================
You should upgrade your Exchange 2013 server roles in the following order:
Client Access
Mailbox
Edge
For DAG members
Upgrade only passive servers in the first phase.
Before you apply the service pack to a DAG member, move all active mailbox database copies off the server to be upgraded and configure the server to be blocked from activation.
Place server in DAG maintenance mode before applying the service pack to any DAG member,
Stop any scheduled tasks or other processes running on the DAG member or within that DAG that could adversely affect the DAG member being upgraded or the upgrade process
Good Luck !!
Remember to mark as helpful if you find my contribution useful or as an answer if it does answer your question.That will encourage me - and others - to take time out to help you Check out my latest blog posts on http://exchangequery.com

DPM Agent Communication Error

We get the following error sporadically, and it seems to always deal with an Exchange Server service stealing the DPMRA port. Is there any way to get this to stop. We have to put the mailbox server in 2010 into DAG maintenance and restart services. This
is not optimal. The below error is what we receive. Seems to be the Microsoft Exchange Replication Service:
DPM failed to communicate with servername because of a communication error with the protection agent. (ID 53 Details: Server execution failed (0x80080005))
We run DPM 2012
Thanks

mkcj84,
What version of DPM are you on> with RU3 for DPM2012R2 you can set what ports DPM uses
http://support.microsoft.com/kb/2966014/en-us
This update lets administrators configure the DPMRA port and select a nondefault port by following these steps:
Install UR3 for DPM 2012 R2.
Run Setagentcfg.exe by using the following command:
setagentcfg.exe s <protected_server_FQDN> <alternative_port>
Notes
In this command, <var><alternative port></var> represents the nondefault port.
By default, this command should be located in the following folder:
%PROGRAMFILES%/Microsoft System Center 2012 R2/DPM/DPM/Setup
Verify that a new entry is created in the following registry subkey:
HKEY_LOCAL_MACHINE \Software\Microsoft\Microsoft Data Protection Manager\Agent\2.0\PsPortConfig
Copy the setagentcfg.exe file from the DPM server to the following folder on the protected computer:
%PROGRAMFILES%\Microsoft Data Protection Manager\DPM\bin
Run the following command on the protected server:
setagentcfg.exe e DPMRA <alternative port>
Note Use the same port number that is specified in step 1.
Restart the DPM server.
Restart the DPMRA service on the protected server.
Regards, A.Nadar, This posting is provided "AS IS" with no warranties, and confers no rights.

DPM Protection Agent Communication Error

We get the following error sporadically, and it seems to always deal with an Exchange Server service stealing the DPMRA port. Is there any way to get this to stop. We have to put the mailbox server in 2010 into DAG maintenance and restart services. This
is not optimal. The below error is what we receive. Seems to be the Microsoft Exchange Replication Service:
DPM failed to communicate with servername because of a communication error with the protection agent. (ID 53 Details: Server execution failed (0x80080005))
We run DPM 2012
Thanks

Hi,
you can tell DPM to use a different port on that exchange server:
http://support2.microsoft.com/default.aspx?scid=kb;EN-US;947682
Cheers
Chaib

After DR fail-back ActiveSync iOS devices still using DR hostname

Have an issue,
Primary Site:
1x CAS
1x MBS
DR Site:
1x CAS/MBX
I performed some maintenance a few nights ago (Windows update and upgrade from 2010 SP3 to RU8v2). Using the DAG maintenance script I moved active DAG copies over to DR and back again.
Since then, ActiveSync on iOS is not working. The production address is 'https://outlook.domain.com' and the DR address is 'https://outlookdr.domain.com'. If I change the iOS devices to point back to production they work fine. Problem is there are a LOT
of these so manually changing is not an option.
Looking at the IIS logs I see:
2015-04-10 09:09:35 10.1.4.24 POST /Microsoft-Server-ActiveSync/default.eas User=User_Name&DeviceId=ApplC39KR5L9DTWD&DeviceType=iPhone&Cmd=Ping&Log=RdirTo:https%3a%2f%2foutlookdr.domain.com%2fMicrosoft-Server-ActiveSync_V141_LdapC1_Cpo20000_Fet20014_Error:MisconfiguredDevice_Mbx:DRServer.domain.local_Budget:(D)Conn%3a1%2cHangingConn%3a0%2cAD%3a%24null%2f%24null%2f0%25%2cCAS%3a%24null%2f%24null%2f0%25%2cAB%3a%24null%2f%24null%2f0%25%2cRPC%3a%24null%2f%24null%2f0%25%2cFC%3a1000%2f0%2cPolicy%3aDefaultThrottlingPolicy%5F88bf084e-f4e4-4e93-bfd9-0285d091bdf0%2cNorm_
443 domain\User_Name 49.224.210.83 Apple-iPhone5C2/1201.365 451 0 0 22058
I'm unsure whats causing iOS devices to keep pulling the wrong information, Androids are fine.
David Robertson

So after a call to premier support, they were stumped. Looks like the initial problem was caused by the DR server having the ActiveSync ExternalURL set.
Currently I have redirected external DNS for 'outlookdr.domain.com' back to 'outlook.domain.com' and clients are receiving email.
Next step is to block the DR firewall NAT and hopefully force clients to autodiscover again to get the correct URL. I'll post with my results.
David Robertson

DAG member, PF database host, and hub maintenance (6hr power-off)

Hello,
We have an EPO test coming up in one of our data centers that requires me to power off 2 of our 4 DAG members. In preparation for this I’ve been reading up on DAG member maintenance best practice as well as deconstructing the Start[Stop]DagServerMaintenance.ps1
script. After testing in my lab I noticed that the iteration of Suspend-MailboxDatabaseCopy is followed by the parameter -ActivationOnly and sets ActivationSuspended to True. Why only suspend copy activation and not database copy processes? Does the script
assume that DAG member maintenance does not include server shutdown or is suspending database copy processes not actually necessary?
Also, one of the DAG members I have to power off serves 1 of 3 of our PF databases. I’ve noticed that the Start[Stop]DagServerMaintenance.ps1 scripts do nothing for PF databases. Should I just unmount it and rely on standard PF replication to update it when
I mount it back? Do I need to change my databases over to use a different PF database?
The power-off duration is 6 hours so I'm not too worried about log build up. Our CAS servers are VMs and will be “motioned” to our other data center. We have two Hub Transport servers, I plan to simply power one of them off. Am I missing anything?
Thanks!

Belinda,
Your understanding and the article you provided on Managing Database Availability Groups seems to be inline with what I've described. However,
just our of curiosity can you or anyone else explain why the script StartDagServerMaintenance.ps1
only runs Suspend-MailboxDatabaseCopy -ActivationOnly? In other words, why does the script only suspend copy activation and not the copy process?
"Did the two Hub Transport servers locate in the same AD site?", yes they are in the same
AD site but different physical locations. The article on HSS helped explain HT HA.
Thanks again!

Taking down DAG member for maintenance

Hello I need some advice on what to do.
I have a two-member Exchange 2010 DAG. One is an active server where everything is mounted. The other is simply the passive server. I will need to power down the passive server in a few days for some extended maintenance. Roughly
several hours. I know for a fact (because I've been through this before) replication will be cut off during this time and as a result I cannot run a full Exchange backup to flush all the transactions logs and secondly, the copy queues will be very backed
up. Last time they were so backed up all efforts to reseed the passive copy of the mailbox databases failed. In the end I think I had to delete and copy the active server's databases over. That took a really really long time.
So to avoid all that headache what should I do for this upcoming maintenance to make life easier?
- Is there a way I can literally STOP replication or even delete the passive copies of the mailbox databases just so I can run full backups and truncate transaction logs?
- Otherwise, assuming I have backed up copy queues, should I simply try reseeding or resuming synchronization? Or should I not even waste time and delete all the passive copies and their files and start fresh with NEW copies?

I've seen the "Failed and Suspended" status and recovered using the Update-MailboxDatabaseCopy
command below. For example:
PS] C:\>Get-MailboxDatabaseCopyStatus DB1\EX13-2 | fl Name,Status,CopyQueueLength,ReplayQueueLength,ContentIndexState
Name : DB1\EX13-2
Status : FailedAndSuspended
CopyQueueLength : 3
ReplayQueueLength : 0
ContentIndexState : Failed
We then update (or reseed) the failed database copy with the data from a good copy:
[PS] C:\>Update-MailboxDatabaseCopy DB1\EX13-2 -DeleteExistingFiles
I understand the reseed process can be lengthy but I'm not sure there are better (supported) methods.
Please mark as helpful if you find my contribution useful or as an answer if it does answer your question. That will encourage me - and others - to take time out to help you.

3 DAG node Maintenance

Hello,
I need to perform maintenance on my 3 DAG node. My primarily is in a Data center in Los Angeles CA, and DR site is in Miami, FL. MB1 holds all the active databases, MB2 and MB3 passive database copies. MB3 is on a DR site.
What is the correct procedure and process to accomplish this with limited user interruptions or creating more issues ?
Please advise,
JOe
My DAG environment
Exchange 2010 SP3 RU2
Windows 2008R2 enterprise
Data center 1 in LA
MB1 Active copy
MB2 Passive Copy
cashub1,cashub2,cashub3
DR Data center 2 in FL
MB3 passive copy.
cashub4

In answer to your questions:
Do you mean, the following steps:
on mb3 mailbox, I run StartDagServerMaintenance.ps1 , performing maintenance, than StopDagServerMaintenance.ps1, reboot mb3 and check all services are running, than follow the same steps for mb2 and mb1? If I misunderstood or if there are
better alternatives, please advise.
Yes. No need to move the databases until you need to patch the server hosting them.
I am not clear as mb3 and mb2 contain passive data base copies, and mb1 has all active database copies. would mb2 and mb3 remain passive, until I do maintenance on mb1 ?
Yes. As I said, why move the databases if you don't need to?
When I perform maintenance on MB1, and running StartDagServerMaintenance.ps1 follow by StartDagServerMaintenance.ps1 when maintenance is completed, will it automatically switch-over to MB2 or MB3? or do I have to do the switch over manually.
I recommend you move the databases manually, and you specify that they be moved to MB2. If they are moved to MB3, all users with mailboxes in those databases will need to exit Outlook and reopen it to get into their mailbox. If you have a CAS
Array in LA, this won't be necessary.

Mailbox Database Size/White Space Reduction in a DAG

I have several large databases in which I'm moving mailboxes off of in an attempt to prevent my LUNs from running out of disk space. As well, I have several mailboxes of users who have left the company. My questions are as follows:
1. When I run remove-mailbox -id "some mailbox" -permanent $true, the mailbox is deleted from Exchange relatively instantly. Will Exchange recoup the white space that is left behind after the mailbox has been deleted so that the
database shrinks in size? Or will it just start writing over the white space rather than growing the size of the database?
2. Perataining to mailboxes that have been moved, if "Keep Deleted Mailboxes for" is set to 4 days on the database that holds mailboxes that I'm moving off, will the database hold onto the mailbox (even though it has successfully been moved to
another database) for the duration of the "Keep Deleted Mailboxes for" setting? Or will the mailbox be hard deleted after the move completes successfully?
3. As part of Online Maintenance, will Exchange defrag a databse and decrease the white space, thereby shrinking the database size or will I have to take it offline to do that?
4. What are the procedures to perform an Offline Defrag of an Exchange Databse? How risky is this?

Hello,
To answer your questions:
1.   When and online database defrag completes a pass on the database it will reclaim this as whitespace in the database. It will not however reclaim space on the disk by shrinking the .edb file. An offline defrag is required for that.
2.   Moved mailbox will not be hard deleted until the 4 day retention period has past, unless you manually purge them using Remove-StoreMailbox.
3.   Related to the answer for number 1. Exchange will fill whitespace as mailboxes grow, but it will not reclaim disk space.
4.   In a DAG you will you need to dismount the mailbox database and run eseutil /d against the active database copy. This is a relatively low risk procedure, but requires databases to be offline thus interrupting mailbox access. It can also
take quite along time depending on the database size. The rate of defrag estimated by MS is 9 GB/hr. Here is a reference:
http://support.microsoft.com/kb/192185/en-us
For these types of situations my preference (If disk space allows) is to move all mailboxes to a fresh database then delete the database with excessive white space

Exchange 2010 DB maintenance

Hello,
Environment : Exchange 2010 sp3 ru3 in DAG
We have a exchange database which is having about 300GB of whitespace.
The database is having about 200 mailboxes and the avaerage mailbox size is 1GB.
We ensured that online maintenance was run completely for 48 hours.We have checked the support sites and see that defrag should not be run on databases which are in DAG.
Last option is to move the mailboxes to another new mailbox.
Any other suggestions?
Regards, Ajit

Agreed with above suggestion.
You can follow the prequisites given at this technet blog and can gather more information while preparing to move mailbox to anoth new mailbox : http://technet.microsoft.com/en-in/library/dd298174%28v=exchg.141%29.aspx
Lepide - Simplifying IT Management

Max. DB Size in DAG

Dear
Although Microsoft supports Exchange 2013 databases up to 16 TB in a DAG, they recommend to use max. 2 TB databases.
Did someone of you already used EX2013 databases larger than 2 TB? What are your experiences?
Thanks for the feedback.
Regards.
Peter
Peter Van Keymeulen, IT Infrastructure Solution Architect, www.edeconsulting.be

Microsoft has recommended a maximum size for an Exchange mailbox database of 2 TB. The current database page size of 32 KB makes the maximum theoretical size of a
database at the NTFS file system limit (16 TB). To be safe, the absolute limit for an Exchange 2010 or Exchange 2013 database is 16,000 GB, just below the NTFS file system theoretical limit.
But In any case, allowing Exchange Server databases to reach multiple terabyte sizes is generally not recommended.
Even though Exchange Server 2013 is perfectly capable of creating and running multi-terabyte databases, using excessively large databases can negatively affect things
like backup, recovery, replication (especially with regard to the initial seeding) and general database maintenance. It's best to limit the maximum after considering above mentioned factors.

Shutting down for maintenance

got 2 mbx nodes in one dag. I will be shutting them down for maintenance. Node A, I will put in maintenance mode and shutdown. whats the best way to shut down nodeB. Should I put that also in maintenance mode bfore I shut down?
What about the cas/ft role server ... I can leave that running I guess?
thanks

Hi,
For the HT/Cas if you are having high availbilty then you can do the maintenance activity then restart the server for mbx if it si dag start by Move-ActiveMailboxDatabase -Server SERVERNAME. similarly for other nodes once done run the script .\RedistributeActiveDatabases.ps1
based upon the activatiion
Regards
Muthu
Thanks Muthu

Rebooting a DAG member in DAC mode

Hi Guys;
I'm about to create a two members DAG with a FSW located in Site A. Both exchange servers are 2010 with SP3. As i 'm just starting my first ever DAG and the lack of knowledge/experience brought me here to avoid any misconfigurations.
I'm planning on a two member DAG, the primary exchange and the FSW servers in site A, in site B a secondary exchange server. I will activate the DAC mode to avoid split brain situation and also will block the automatic activation of databases on the server
in Site B, so i can have a full control over it, as power problems are a bit concern (it happens from time to time)
Now what confuses me, is the windows update or exchange related patch installations.
I would like to know if i can do regular windows updates in this scenario without moving the databases to the secondary server, and so for the exchange related updates.
What is the best approach in this case?
Thanks in advance.

Well, I wouldn't set auto activation to block in this case so that whenever something happens to databases (for example putting server in Site A in maintenance for Windows or Exchange patching), all of them will be mounted on server in site B as next preference
and users won't see any impact.
If power goes for both servers in Site A (Exchange and witness) then database will be in service down situation since you have DAC mode setup and you need to execute the DAC mode activation anyways to mount all the database on Site B manually taking decision....
http://eightwone.com/2010/08/30/datacenter-activation-coordination-mode/

Shut down DAG gracefully

I have a practice DAG network of two Exchange 2010 (SP3) servers and a FSW on a file server.
What is the best way to stop the entire DAG, gracefully, without causing problems the next time I want to use it?
In other words, since this is for practice (and does not really need to provide high availability for any organization), how can I prevent a failover from taking place when I shut down the first server?
How can I prevent the nodes from being confused when I turn my network back on?
Stop the cluster service on the two nodes?
Run the Stop-DatabaseAvailabilityGroup cmdlet first?
I realize this is not typical since, normally, you would always want at least one node available.
I could leave my machines running all the time, but that would waste energy when I'm not practicing various operations.
Thanks in advance!
Please mark as helpful if you find my contribution useful or as an answer if it does answer your question. That will encourage me - and others - to take time out to help you.

Sounds like it is your test lab... I wouldn't worry about shutdown sequence or procedure for DAG in Test environment... In fact I would love to mess it up with not doing that...
Regardless you can put a server into maintenance mode and shut it down so even if you bring that server up first, it won't mount the database because maintenance mode script sets the DatabaseCopyAutoActivationPolicy to Blocked on that server.
You can shutdown the second server after that...

DAG maintenance

Similar Messages

Maybe you are looking for