| Author | Messages | |
Marty1_0
Posts:0
 | | 07/11/2009 9:14 AM |
| Hi,
I was wondering if there were general guidelines for maintenance on Exchange
due a restore I have done recently. Recovery of the DB took me 28 hours
(47GB file) before I was able to mount it.
- how often do an offline defragementation? Which criteria do count to
determine this?
- what are general mailbox quota, even if you don't have size restrictions
on server side (not that many users, disk space enough)?
...
Thanks for all opinions and guidelines
-Bart
| | | |
| michael1
Posts:426
| | Marty1_0
Posts:0
| | 07/12/2009 9:29 AM |
| Well they did a restore of their mailstore and logfiles, but they weren't
able to mount it due inconsistent state. When I check the store, it was in
state 'Dirty Shutdown'. Finally I ran a recovery on it which took approx. 28
hours to finish. Then I was able to mount the store. This is an Exchange
2003 server.
Next week I will check their configuration to see if they have online
defragmentation configured. Also a collegue of mine will reconfigure their
backups next week as they don't have GRT backups of Exchange and they want
this.
This started me wondering on maintenance guidelines. Side note: Exchange is
a VM, no idea what the impact is?
Thanks, I will go through the links given :-)
-Bart
On Sat, Jul 11, 2009 at 3:34 PM, Michael B. Smith <
michael@theessentialexchange.com> wrote:
> Why on earth did it take you 28 hours and what caused it? I would be
> interested in the time line of that event.
>
>
>
> For backups on Exchange 2000 and if you are doing streaming backups on
> Exchange 2003, you should limit your Exchange stores to around 50 GB each –
> for most people this allows you to meet a standard recovery time objective
> of four (yes, 4) hours.
> http://technet.microsoft.com/en-us/library/aa995715.aspx
>
>
>
> If you are doing VSS backups on Exchange 2003, or you are on Exchange 2007
> without using one of the continuous replication (CR) mechanisms, you should
> limit your Exchange stores to around 100 GB each. If you are using a CR
> solution on Exchange 2007 or above, you should limit your stores to around
> 200 GB each (difference being that you have a copy you can “immediately”
> fail-over to). http://technet.microsoft.com/en-us/library/bb124704.aspx
>
>
>
> All that being said, these are only guidelines. It’s dependent on your
> hardware, your backup and restore windows, your recovery time objective, and
> your experience in performing these operations (and yes, you should PRACTICE
> them). There are no known technical reasons to prevent an Exchange store
> approaching the architectural limit (8 TB and above, depending on version)
> and to my personal dismay I’ve worked with mailstores almost at 2 TB before.
>
>
>
> Offline defragmentation is never recommended as a normal maintenance
> method. http://msexchangeteam.com/archive/2004/07/08/177574.aspx
>
>
>
> There are rare places and situations where it makes sense (educational
> institutions at the end of a semester, for example), but there are almost
> always better solutions that require no downtime (moving the remaining
> contents of a mailstore to a new mailstore and then deleting the old
> mailstore, for example).
>
>
>
> Online defragmentation should complete at least once a week per store. (I
> didn’t look up an URL for that, you can find one if you want one.) JPersonally, I prefer online defragmentation to complete every night.
>
>
>
> I just finished deploying a client this week who has about 1,500 mailboxes.
> They were space-limited on their old solution (200 GB on a SAN) and had
> mailbox quotas of 75 MB per mailbox, but they are not space limited on the
> new one (6 TB of RAID-10 DAS). I recommended 2 GB limits (which is the
> default limit in Exchange 2007). Even if all of their mailboxes hit the
> quota, they won’t run out of disk space…(they never wanted to be space
> limited again – the solution is a bit of overkill in that regard). Setting
> SOME limit is administratively a good idea, simply to prevent a
> mailbox/mailuser gone awry – such as the mail loop as we experienced here a
> few weeks ago.
>
>
>
> Exchange 2007 sp1 and above with Outlook 2007 sp2 and above can easily
> handle 5 GB mailboxes and I’ve got clients with mailboxes that exceed 10 GB.
> As long as your planning and monitoring are good, “the sky is the limit”.
>
>
>
> There are consultants and companies that do Exchange Health Checks, just as
> there are folks who do Active Directory Health Checks.
>
>
>
> The deployment and planning guides on TechNet for Exchange 2007 and above
> are very good.
>
>
>
> Regards,
>
>
>
> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>
> My blog: http://TheEssentialExchange.com/blogs/michael<http://theessentialexchange.com/blogs/michael>
>
> Monitoring Exchange w/OpsMgr now available http://snurl.com/45ppf
>
>
>
> *From:* activedir-owner@mail.activedir.org [mailto:
> activedir-owner@mail.activedir.org] *On Behalf Of *Bart Van den Wyngaert
> *Sent:* Saturday, July 11, 2009 4:12 AM
> *To:* ActiveDir
> *Subject:* [ActiveDir] OT: Exchange maintenance
>
>
>
> Hi,
>
>
>
> I was wondering if there were general guidelines for maintenance on
> Exchange due a restore I have done recently. Recovery of the DB took me 28
> hours (47GB file) before I was able to mount it.
>
> - how often do an offline defragementation? Which criteria do count to
> determine this?
>
> - what are general mailbox quota, even if you don't have size restrictions
> on server side (not that many users, disk space enough)?
>
> ...
>
>
>
> Thanks for all opinions and guidelines
>
> -Bart
>
| | | |
| andrew
Posts:77
| | 07/12/2009 12:55 PM |
| May I also take this opportunity to thank Michael for his excellent and
concise answer.
2009/7/12 Bart Van den Wyngaert <bart.vdw@gmail.com>
> Well they did a restore of their mailstore and logfiles, but they weren't
> able to mount it due inconsistent state. When I check the store, it was in
> state 'Dirty Shutdown'. Finally I ran a recovery on it which took approx. 28
> hours to finish. Then I was able to mount the store. This is an Exchange
> 2003 server.
>
> Next week I will check their configuration to see if they have online
> defragmentation configured. Also a collegue of mine will reconfigure their
> backups next week as they don't have GRT backups of Exchange and they want
> this.
>
> This started me wondering on maintenance guidelines. Side note: Exchange is
> a VM, no idea what the impact is?
>
> Thanks, I will go through the links given :-)
>
> -Bart
>
> On Sat, Jul 11, 2009 at 3:34 PM, Michael B. Smith <
> michael@theessentialexchange.com> wrote:
>
>> Why on earth did it take you 28 hours and what caused it? I would be
>> interested in the time line of that event.
>>
>>
>>
>> For backups on Exchange 2000 and if you are doing streaming backups on
>> Exchange 2003, you should limit your Exchange stores to around 50 GB each –
>> for most people this allows you to meet a standard recovery time objective
>> of four (yes, 4) hours.
>> http://technet.microsoft.com/en-us/library/aa995715.aspx
>>
>>
>>
>> If you are doing VSS backups on Exchange 2003, or you are on Exchange 2007
>> without using one of the continuous replication (CR) mechanisms, you should
>> limit your Exchange stores to around 100 GB each. If you are using a CR
>> solution on Exchange 2007 or above, you should limit your stores to around
>> 200 GB each (difference being that you have a copy you can “immediately”
>> fail-over to). http://technet.microsoft.com/en-us/library/bb124704.aspx
>>
>>
>>
>> All that being said, these are only guidelines. It’s dependent on your
>> hardware, your backup and restore windows, your recovery time objective, and
>> your experience in performing these operations (and yes, you should PRACTICE
>> them). There are no known technical reasons to prevent an Exchange store
>> approaching the architectural limit (8 TB and above, depending on version)
>> and to my personal dismay I’ve worked with mailstores almost at 2 TB before.
>>
>>
>>
>> Offline defragmentation is never recommended as a normal maintenance
>> method. http://msexchangeteam.com/archive/2004/07/08/177574.aspx
>>
>>
>>
>> There are rare places and situations where it makes sense (educational
>> institutions at the end of a semester, for example), but there are almost
>> always better solutions that require no downtime (moving the remaining
>> contents of a mailstore to a new mailstore and then deleting the old
>> mailstore, for example).
>>
>>
>>
>> Online defragmentation should complete at least once a week per store. (I
>> didn’t look up an URL for that, you can find one if you want one.) JPersonally, I prefer online defragmentation to complete every night.
>>
>>
>>
>> I just finished deploying a client this week who has about 1,500
>> mailboxes. They were space-limited on their old solution (200 GB on a SAN)
>> and had mailbox quotas of 75 MB per mailbox, but they are not space limited
>> on the new one (6 TB of RAID-10 DAS). I recommended 2 GB limits (which is
>> the default limit in Exchange 2007). Even if all of their mailboxes hit the
>> quota, they won’t run out of disk space…(they never wanted to be space
>> limited again – the solution is a bit of overkill in that regard). Setting
>> SOME limit is administratively a good idea, simply to prevent a
>> mailbox/mailuser gone awry – such as the mail loop as we experienced here a
>> few weeks ago.
>>
>>
>>
>> Exchange 2007 sp1 and above with Outlook 2007 sp2 and above can easily
>> handle 5 GB mailboxes and I’ve got clients with mailboxes that exceed 10 GB.
>> As long as your planning and monitoring are good, “the sky is the limit”.
>>
>>
>>
>> There are consultants and companies that do Exchange Health Checks, just
>> as there are folks who do Active Directory Health Checks.
>>
>>
>>
>> The deployment and planning guides on TechNet for Exchange 2007 and above
>> are very good.
>>
>>
>>
>> Regards,
>>
>>
>>
>> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>>
>> My blog: http://TheEssentialExchange.com/blogs/michael<http://theessentialexchange.com/blogs/michael>
>>
>> Monitoring Exchange w/OpsMgr now available http://snurl.com/45ppf
>>
>>
>>
>> *From:* activedir-owner@mail.activedir.org [mailto:
>> activedir-owner@mail.activedir.org] *On Behalf Of *Bart Van den Wyngaert
>> *Sent:* Saturday, July 11, 2009 4:12 AM
>> *To:* ActiveDir
>> *Subject:* [ActiveDir] OT: Exchange maintenance
>>
>>
>>
>> Hi,
>>
>>
>>
>> I was wondering if there were general guidelines for maintenance on
>> Exchange due a restore I have done recently. Recovery of the DB took me 28
>> hours (47GB file) before I was able to mount it.
>>
>> - how often do an offline defragementation? Which criteria do count to
>> determine this?
>>
>> - what are general mailbox quota, even if you don't have size restrictions
>> on server side (not that many users, disk space enough)?
>>
>> ...
>>
>>
>>
>> Thanks for all opinions and guidelines
>>
>> -Bart
>>
>
>
| | | |
| jamesawells
Posts:79
| | 07/12/2009 1:29 PM |
| I would guess that the extraordinary amount of time that the recovery
took might be related to storage performance, especially since this is
on a VM.
I'd imagine that the Physical Disk\Avg sec/Write and Avg sec/Read were
both way to high on the DB and Log drives...
--James
On Sun, Jul 12, 2009 at 4:27 AM, Bart Van den
Wyngaert<bart.vdw@gmail.com> wrote:
> Well they did a restore of their mailstore and logfiles, but they weren't
> able to mount it due inconsistent state. When I check the store, it was in
> state 'Dirty Shutdown'. Finally I ran a recovery on it which took approx. 28
> hours to finish. Then I was able to mount the store. This is an Exchange
> 2003 server.
>
> Next week I will check their configuration to see if they have online
> defragmentation configured. Also a collegue of mine will reconfigure their
> backups next week as they don't have GRT backups of Exchange and they want
> this.
>
> This started me wondering on maintenance guidelines. Side note: Exchange is
> a VM, no idea what the impact is?
>
> Thanks, I will go through the links given :-)
>
> -Bart
>
> On Sat, Jul 11, 2009 at 3:34 PM, Michael B. Smith
> <michael@theessentialexchange.com> wrote:
>>
>> Why on earth did it take you 28 hours and what caused it? I would be
>> interested in the time line of that event.
>>
>>
>>
>> For backups on Exchange 2000 and if you are doing streaming backups on
>> Exchange 2003, you should limit your Exchange stores to around 50 GB each –
>> for most people this allows you to meet a standard recovery time objective
>> of four (yes, 4) hours.
>> http://technet.microsoft.com/en-us/library/aa995715.aspx
>>
>>
>>
>> If you are doing VSS backups on Exchange 2003, or you are on Exchange 2007
>> without using one of the continuous replication (CR) mechanisms, you should
>> limit your Exchange stores to around 100 GB each. If you are using a CR
>> solution on Exchange 2007 or above, you should limit your stores to around
>> 200 GB each (difference being that you have a copy you can “immediately”
>> fail-over to). http://technet.microsoft.com/en-us/library/bb124704.aspx
>>
>>
>>
>> All that being said, these are only guidelines. It’s dependent on your
>> hardware, your backup and restore windows, your recovery time objective, and
>> your experience in performing these operations (and yes, you should PRACTICE
>> them). There are no known technical reasons to prevent an Exchange store
>> approaching the architectural limit (8 TB and above, depending on version)
>> and to my personal dismay I’ve worked with mailstores almost at 2 TB before.
>>
>>
>>
>> Offline defragmentation is never recommended as a normal maintenance
>> method. http://msexchangeteam.com/archive/2004/07/08/177574.aspx
>>
>>
>>
>> There are rare places and situations where it makes sense (educational
>> institutions at the end of a semester, for example), but there are almost
>> always better solutions that require no downtime (moving the remaining
>> contents of a mailstore to a new mailstore and then deleting the old
>> mailstore, for example).
>>
>>
>>
>> Online defragmentation should complete at least once a week per store. (I
>> didn’t look up an URL for that, you can find one if you want one.) J
>> Personally, I prefer online defragmentation to complete every night.
>>
>>
>>
>> I just finished deploying a client this week who has about 1,500
>> mailboxes. They were space-limited on their old solution (200 GB on a SAN)
>> and had mailbox quotas of 75 MB per mailbox, but they are not space limited
>> on the new one (6 TB of RAID-10 DAS). I recommended 2 GB limits (which is
>> the default limit in Exchange 2007). Even if all of their mailboxes hit the
>> quota, they won’t run out of disk space…(they never wanted to be space
>> limited again – the solution is a bit of overkill in that regard). Setting
>> SOME limit is administratively a good idea, simply to prevent a
>> mailbox/mailuser gone awry – such as the mail loop as we experienced here a
>> few weeks ago.
>>
>>
>>
>> Exchange 2007 sp1 and above with Outlook 2007 sp2 and above can easily
>> handle 5 GB mailboxes and I’ve got clients with mailboxes that exceed 10 GB.
>> As long as your planning and monitoring are good, “the sky is the limit”.
>>
>>
>>
>> There are consultants and companies that do Exchange Health Checks, just
>> as there are folks who do Active Directory Health Checks.
>>
>>
>>
>> The deployment and planning guides on TechNet for Exchange 2007 and above
>> are very good.
>>
>>
>>
>> Regards,
>>
>>
>>
>> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
>>
>> My blog: http://TheEssentialExchange.com/blogs/michael
>>
>> Monitoring Exchange w/OpsMgr now available http://snurl.com/45ppf
>>
>>
>>
>> From: activedir-owner@mail.activedir.org
>> [mailto:activedir-owner@mail.activedir.org] On Behalf Of Bart Van den
>> Wyngaert
>> Sent: Saturday, July 11, 2009 4:12 AM
>> To: ActiveDir
>> Subject: [ActiveDir] OT: Exchange maintenance
>>
>>
>>
>> Hi,
>>
>>
>>
>> I was wondering if there were general guidelines for maintenance on
>> Exchange due a restore I have done recently. Recovery of the DB took me 28
>> hours (47GB file) before I was able to mount it.
>>
>> - how often do an offline defragementation? Which criteria do count to
>> determine this?
>>
>> - what are general mailbox quota, even if you don't have size restrictions
>> on server side (not that many users, disk space enough)?
>>
>> ...
>>
>>
>>
>> Thanks for all opinions and guidelines
>>
>> -Bart
>
| | | |
| Marty1_0
Posts:0
| | 07/12/2009 1:50 PM |
| I have to monitor that to be sure... thanks
-Bart
On Sun, Jul 12, 2009 at 2:29 PM, James Wells <jamwel@gmail.com> wrote:
> I would guess that the extraordinary amount of time that the recovery
> took might be related to storage performance, especially since this is
> on a VM.
>
> I'd imagine that the Physical Disk\Avg sec/Write and Avg sec/Read were
> both way to high on the DB and Log drives...
>
>
> --James
>
> On Sun, Jul 12, 2009 at 4:27 AM, Bart Van den
> Wyngaert<bart.vdw@gmail.com> wrote:
> > Well they did a restore of their mailstore and logfiles, but they weren't
> > able to mount it due inconsistent state. When I check the store, it was
> in
> > state 'Dirty Shutdown'. Finally I ran a recovery on it which took approx.
> 28
> > hours to finish. Then I was able to mount the store. This is an Exchange
> > 2003 server.
> >
> > Next week I will check their configuration to see if they have online
> > defragmentation configured. Also a collegue of mine will reconfigure
> their
> > backups next week as they don't have GRT backups of Exchange and they
> want
> > this.
> >
> > This started me wondering on maintenance guidelines. Side note: Exchange
> is
> > a VM, no idea what the impact is?
> >
> > Thanks, I will go through the links given :-)
> >
> > -Bart
> >
> > On Sat, Jul 11, 2009 at 3:34 PM, Michael B. Smith
> > <michael@theessentialexchange.com> wrote:
> >>
> >> Why on earth did it take you 28 hours and what caused it? I would be
> >> interested in the time line of that event.
> >>
> >>
> >>
> >> For backups on Exchange 2000 and if you are doing streaming backups on
> >> Exchange 2003, you should limit your Exchange stores to around 50 GB
> each –
> >> for most people this allows you to meet a standard recovery time
> objective
> >> of four (yes, 4) hours.
> >> http://technet.microsoft.com/en-us/library/aa995715.aspx
> >>
> >>
> >>
> >> If you are doing VSS backups on Exchange 2003, or you are on Exchange
> 2007
> >> without using one of the continuous replication (CR) mechanisms, you
> should
> >> limit your Exchange stores to around 100 GB each. If you are using a CR
> >> solution on Exchange 2007 or above, you should limit your stores to
> around
> >> 200 GB each (difference being that you have a copy you can “immediately”
> >> fail-over to). http://technet.microsoft.com/en-us/library/bb124704.aspx
> >>
> >>
> >>
> >> All that being said, these are only guidelines. It’s dependent on your
> >> hardware, your backup and restore windows, your recovery time objective,
> and
> >> your experience in performing these operations (and yes, you should
> PRACTICE
> >> them). There are no known technical reasons to prevent an Exchange store
> >> approaching the architectural limit (8 TB and above, depending on
> version)
> >> and to my personal dismay I’ve worked with mailstores almost at 2 TB
> before.
> >>
> >>
> >>
> >> Offline defragmentation is never recommended as a normal maintenance
> >> method. http://msexchangeteam.com/archive/2004/07/08/177574.aspx
> >>
> >>
> >>
> >> There are rare places and situations where it makes sense (educational
> >> institutions at the end of a semester, for example), but there are
> almost
> >> always better solutions that require no downtime (moving the remaining
> >> contents of a mailstore to a new mailstore and then deleting the old
> >> mailstore, for example).
> >>
> >>
> >>
> >> Online defragmentation should complete at least once a week per store.
> (I
> >> didn’t look up an URL for that, you can find one if you want one.) J
> >> Personally, I prefer online defragmentation to complete every night.
> >>
> >>
> >>
> >> I just finished deploying a client this week who has about 1,500
> >> mailboxes. They were space-limited on their old solution (200 GB on a
> SAN)
> >> and had mailbox quotas of 75 MB per mailbox, but they are not space
> limited
> >> on the new one (6 TB of RAID-10 DAS). I recommended 2 GB limits (which
> is
> >> the default limit in Exchange 2007). Even if all of their mailboxes hit
> the
> >> quota, they won’t run out of disk space…(they never wanted to be space
> >> limited again – the solution is a bit of overkill in that regard).
> Setting
> >> SOME limit is administratively a good idea, simply to prevent a
> >> mailbox/mailuser gone awry – such as the mail loop as we experienced
> here a
> >> few weeks ago.
> >>
> >>
> >>
> >> Exchange 2007 sp1 and above with Outlook 2007 sp2 and above can easily
> >> handle 5 GB mailboxes and I’ve got clients with mailboxes that exceed 10
> GB.
> >> As long as your planning and monitoring are good, “the sky is the
> limit”.
> >>
> >>
> >>
> >> There are consultants and companies that do Exchange Health Checks, just
> >> as there are folks who do Active Directory Health Checks.
> >>
> >>
> >>
> >> The deployment and planning guides on TechNet for Exchange 2007 and
> above
> >> are very good.
> >>
> >>
> >>
> >> Regards,
> >>
> >>
> >>
> >> Michael B. Smith, MCITP:SA,EMA/MCSE/Exchange MVP
> >>
> >> My blog: http://TheEssentialExchange.com/blogs/michael<http://theessentialexchange.com/blogs/michael>
> >>
> >> Monitoring Exchange w/OpsMgr now available http://snurl.com/45ppf
> >>
> >>
> >>
> >> From: activedir-owner@mail.activedir.org
> >> [mailto:activedir-owner@mail.activedir.org] On Behalf Of Bart Van den
> >> Wyngaert
> >> Sent: Saturday, July 11, 2009 4:12 AM
> >> To: ActiveDir
> >> Subject: [ActiveDir] OT: Exchange maintenance
> >>
> >>
> >>
> >> Hi,
> >>
> >>
> >>
> >> I was wondering if there were general guidelines for maintenance on
> >> Exchange due a restore I have done recently. Recovery of the DB took me
> 28
> >> hours (47GB file) before I was able to mount it.
> >>
> >> - how often do an offline defragementation? Which criteria do count to
> >> determine this?
> >>
> >> - what are general mailbox quota, even if you don't have size
> restrictions
> >> on server side (not that many users, disk space enough)?
> >>
> >> ...
> >>
> >>
> >>
> >> Thanks for all opinions and guidelines
> >>
> >> -Bart
> >
>
| | | |
| michael1
Posts:426
| |
|