Re: Out of memory?
- From: Curt <Curt@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Sun, 26 Apr 2009 23:00:01 -0700
Dell PowerEdge 2800 with Perc 4 controller, SCSI disks, RAID 5. (2
volumes...I think both are raid 5)
I have no disk/raid errors.
"kj [SBS MVP]" wrote:
Diags are usually pretty good about memory and cpu test and the disk.
controller to a large degree.
Extensive disk testing needs to read / write to the disk and aren't
comprehsive without being destructive. Having so many database problems
would make me suspecious of a disk. Even a decent disk controller should be
reporting errors - good ones would be throwing major errors.
What kind of disk, raid controller, raid configuration do you have?
Curt wrote:
I reboot to Dell diagnostics and they all pass...will verify again.
I have issues with wins db and SQL db (specifically backupexec db).
I want to try restore from Thursday full backup (crashed Friday
morning), but I cannot get MSSQL$BKUPEXEC to start...and hence
backupexec fails to start.
Strange thing keeps happening. In directory services respair mode,
explorer keeps crashing.
...an indication of disk/controller/memory issues?
"kj [SBS MVP]" wrote:
Your DB is corrupt, either on disk or durring read from disk and
transfer to memory. Most likely the former. The DB is cached as it
is read into memory but it can exceed the available memory. It will
perform *much* better if it can be loaded into memory, but it
doesn't have to all fit. Fragmentation and white space again are
performance issues, not integrity ones.
The errors you are getting indicate data integrity problems
(corruption) and not memory / fragmentation / page file problems.
You should be looking at doing DB restore and verify the
disk/controller if you haven't already.
Curt wrote:
3.2GB free on c now
Tried ntdsutil Checksum....errors galore....going to try compact.
3330 pages seen
2924 bad checksums
404 uninitialized pages
2924 wrong page numbers
0x0 highest dbtime (pgno 0x0)
417 reads performed
26 MB read
2 seconds taken
13 MB/second
459213 milliseconds used
1101 milliseconds per read
2015 milliseconds for the slowest read
31 milliseconds for the fastest read
Operation terminated with error -1206 (JET_errDatabaseCorrupted, Non
database fi
le or corrupted db) after 2.94 seconds.
Spawned Process Exit code 0xfffffb4a(-1206)
"Cliff Galiher" wrote:
2 gigs is not much on a 12 gig drive. Clean up more. Try to get
to *at least* 25%.
Is your swapfile on the C drive?
I suspect your AD got corrupt in a bad way (I recall your previous
thread) and now is larger than your system can handle because of
the corruption. The fact that ntdsutil *still* can't load your
database only reaffirms this. ntdsutil does not use caching by
design.
Check this out and see if it helps:
http://support.microsoft.com/kb/948603
I've never seen it on SBS before because SBS's 75 user limit
*usually* prevents AD from getting so big that this surfaces...but
a corrupt AD db is an entirely different story. Follow the
procedure and see if you can get an offline defrag going...
-Cliff
"Curt" <Curt@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:D85FC6AA-0C9D-42C5-8A4E-C257EE877195@xxxxxxxxxxxxxxxx
By the way...best I can tell, the server was downloading windows
updates (set
to download but not install) when problem occured.
Could have filled the C drive (12GB partition with 2Gb free after
I cleaned
up a bit).
...no smoking gun, but analysis not as important as fixing right
now.
"Curt" wrote:
SBS2003 with 4 GB memory 4095 virtual memory setting
Before reboot lots of event ID 17055
17052 :
A floating point exception occurred in the user process. Current
transaction
is canceled.
Client reboot as could not log into domain.
Directory services could not start (I suspected corrupt ntds.dit
database.
Try to check integrity with ntdsutil get error -1011
jet_erroutofmemory. Eventually used eseutil to check
database/recover. Still won't start directory services
wins database error -1011 as well.
...so I am thinking that the server has memory issue.
I have rebooted to Dell diagnostic partition and all memory and
system tests
pass.
I removed ram and booted with each one individually...no
difference.
I see 3.25 GB memory...which is correct.
In directory services repair mode, how can I test available
memory? I look at taskmgr and I don't see all memory used, but
when I boot normally,
I cannot get directory services to load (it errors with error
-1011 and cannot load the ntds database).
--
/kj
--
/kj
- Follow-Ups:
- Re: Out of memory?
- From: kj [SBS MVP]
- Re: Out of memory?
- References:
- Out of memory?
- From: Curt
- RE: Out of memory?
- From: Curt
- Re: Out of memory?
- From: Cliff Galiher
- Re: Out of memory?
- From: Curt
- Re: Out of memory?
- From: kj [SBS MVP]
- Re: Out of memory?
- From: Curt
- Re: Out of memory?
- From: kj [SBS MVP]
- Out of memory?
- Prev by Date: Re: Out of memory?
- Next by Date: Re: Out of memory?
- Previous by thread: Re: Out of memory?
- Next by thread: Re: Out of memory?
- Index(es):
Relevant Pages
|
Loading
