Re: Issues with Store.exe dying during Backup (using Backup Exec 10)
- From: "andy webb" <awebb@xxxxxxxxxxxxxxxxxxxxxxx>
- Date: Sat, 4 Mar 2006 21:27:50 -0600
....or stop doing brick backups. ;)
great answer, btw.
"John Fullbright" <fullbrij@xxxxxxxxxxx> wrote in message
news:OujtkS7PGHA.1728@xxxxxxxxxxxxxxxxxxxxxxx
Exchmem does lots of things, one of them is to manages heaps. An exchange
database is basically a big mapped file. Of course the whole thing won't
fit into RAM, so eachange maps pages into RAM, more specifically the
database buffers. When a requested page is not present, a page fault
occurs and the page is retrieved from disk. When buffers in RAM become
full, an LRU-K algorithm is used to determine which pages to drop.
Exchange dynamically allocates buffers using up to 80% or so of available
RAM. On a system with more than 2G of RAM and the /3gb switch is present
in the boot.ini, 898MB of RAM is used for the database buffers. Log
buffers are also allocated based on the setting of the
msexchESEParamLogBuffers attribute for each storage group. Other exchange
processes also use memory, motst notably IIS. Then, of course other user
mode processes on the box, and system processes.
Usually, when I see errors in Exchmem, it's a situation where memory is
constrained. In this case, new pages cannot load into the database
buffers until old pages are trimmed or dropped. That the problem occurs
during a brick level backup would make sense in this scenario; we're
reading lots of pages as we go through each mailbox doing a brick level
backup.
I'd start by making sure we don't have a resource shortage.
How much ram is in the system?
How many storage groups are there?
How many log buffers per storage group?
How may database buffers?
What is the total memory used by store.exe?
Next check a few counters. Database - Page fault stalls/sec and Database
log record stalls/sec, plus the normal memory counters. It may be a case
where you just need to add more RAM if you are not already at 4GB.
"YDKCooKiE" <funk@xxxxxxx> wrote in message
news:pejj02tq9a40mlggv4f7crlib9646vbvhg@xxxxxxxxxx
When I'm backing up my users' mailboxes, I've been having an issue in
where Backup Exec 10 is losing connection to the Exchange Server
(2003, SP2) when it's backing up because store.exe is dying. Here's
the event logs that appear when it dies:
Faulting application store.exe, version 6.5.7638.2, faulting module
exchmem.dll, version 6.5.7638.1, fault address 0x000053b4.
- Application error (1000)
Faulting application store.exe, version 6.5.7638.2, stamp 434353c1,
faulting module exchmem.dll, version 6.5.7638.1, stamp 430e813e,
debug? 0, fault address 0x000053b4.
- Microsoft Exchange Serv (1000)
I'm hoping someone might have come across this or might have any
suggestions as to what might be causing this? Any help is
appriciated!
Thanks,
Curt L.
.
- References:
- Issues with Store.exe dying during Backup (using Backup Exec 10)
- From: YDKCooKiE
- Re: Issues with Store.exe dying during Backup (using Backup Exec 10)
- From: John Fullbright
- Issues with Store.exe dying during Backup (using Backup Exec 10)
- Prev by Date: Re: w3wp.exe filling up Windows\Temp
- Next by Date: Re: why isn't email reaching recipients who are outside domain.
- Previous by thread: Re: Issues with Store.exe dying during Backup (using Backup Exec 10)
- Next by thread: OtherMailbox property info
- Index(es):
Relevant Pages
|
