RE: Information Store Corrupted
- From: David Ruth <DavidRuth@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Wed, 25 Apr 2007 06:34:03 -0700
Obviously something disrupted a disk write. The somethings that are possible
depend on the kind of storage you are using. Direct Attached Storage is the
simplest and various SAN technologies are more complex and have more ways for
it to get disrupted and therefore more difficult to troubleshoot. iSCSI can
be fun when tracking these issues down. Get some vendor support if you need
it. You don't want to make this a habit.
Another potential solution not involving downtime but with the prospect of
losing the already corrupted data is simply move all the mailboxes to a new
storage group (same or new server as appropriate) setting "Skipped corrupted
items" to 1. If the corruption is a database structure page you will not
loss any user data. If the corruption is user data then you will loss that
user data because it will be left behind. I am not recommending it because
it is never a best practice to sacrifice user data but sometimes you need a
different approach. Everything we do involves tradeoffs. This is another
approach for your toolkit. You can, of course, adjust the level of
acceptable data loss by adjusting the "skip corupted items" to a higher
number.
Oh by the way... You discarded the corrupt pages by using the eseutil /p
switch. Given your end state you would have been better off moving the
mailboxes with "Skip corrupted items" and avoiding the downtime.
Others have explained the Best Practise approach of restoring from tape and
running logs. The backup by definition doesn't include the corrupted items.
I hope this helps.
"Shanthi" wrote:
I am using Exchange 2003 on Win2K3 server and backup application is Veritas..
My backup method is for
Daily - differential backup of Information Stores
Weekly - full backup of Information Stores
When I tried to run weekly backup an error has come as "Information Store -
is a corrupt file. This file cannot verify."
I checked the event log on Exchange server, it threw application error with
Source : ESE
Category : Database Page Cache
Event Id : 474
Description :
Information Store (2712) First Storage Group: The database page read from
the file "E:\Exchange\mdbdata\priv1.edb" at offset 409464832
(0x000000001867f000) (database page 99966 (0x1867E)) for 4096 (0x00001000)
bytes failed verification due to a page checksum mismatch. The expected
checksum was 3567500155055337822 (0x31824e7d5639715e) and the actual checksum
was 3711585678364162506 (0x338233827852b9ca). The read operation will fail
with error -1018 (0xfffffc06). If this condition persists then please
restore the database from a previous backup. This problem is likely due to
faulty hardware. Please contact your hardware vendor for further assistance
diagnosing the problem.
I tried some search on Microsoft online technical articles and found useful
informations. Below are the activities performed to resolve the issue;
Step 1: Taken a copy of all raw database.
Step 2: Run eseutil /p for repair database.
Step 3: Taken a backup of repaired database
Step 4: Run eseutil /d for defragment database.
Step 5: Taken a backup of defragmented database.
Step 6: Run Isinteg -fix -test alltests for checking integrity.
Step 7: Found no error in integrity checking.
Step 8: Run ntbackup on exchange server and completed with no issues.
Step 9: Tried full backup of information store by using veritas backup exec
and completed succesfully.
Please do let me know whether the way I performed is correct and also why
this kind of errors are coming. I kept my mail users on waiting for 26 hours
(such a huge down time)
Shanthi
- Prev by Date: Re: Move Internet mail service within same site - exchange 2003
- Next by Date: Exchange 2007 DL lists issue universal\domain local\global
- Previous by thread: Re: Information Store Corrupted
- Next by thread: Exchange 2007 forest prep
- Index(es):
Relevant Pages
|
Loading
