We are experiencing the same errors and same "lock out" behavior. We have th
e same version/patch level of SQL installed but are in a 32 bit env. The box
has 32GB of memory. Nether the bPool or VAS show any pressure when the "out
of system memory error" oc
curs. Since this error usually indicates a VAS problem we tried increasing V
AS to 1GB w/the -g flag. It made no difference. PSS has been working on the
case for 3 weeks. They dont seem to be finding any evidince of memory pressu
re either. When I last spol
e to the escalation engineer yesterday it seemed that they are going to reco
mmend reducing the number of databases on the server. I asked for clarificat
ion as to whether we are hitting a 32 bit barrior, an instance limitation, o
r both. I am awaiting the a
nswer. How many databases do you have on your server? We had between 1700 an
d 1900 (the number varies) at times when the error occured. We are now at 15
00, and have not had the error in the 2 days since reducing the number of da
tabases...
EggHeadCafe.com - .NET Developer Portal of Choice
http://www.eggheadcafe.comSteve,
Further comment.
I have been having the problem about once a month for several months and
have not had any joy in discovering what to do about it. Other notes on my
case:
- Once this starts, my CPU utilization will go to very close to 100%. If I
log out some heavy users, my CPU will dip to 70% or 80% for about 20
seconds, then climb back to near 100%.
- At this point, spid 1 is the busiest of the spids. (Which makes it seem
that the SQL Server is eating up its resources trying to find resources for
users.)
- Very Strange Discovery. One time I found that one workstation which had
been running a very large transaction, would, if turned off, drop the server
CPU to normal. But if I turned that workstation back on, even though I did
not access the server, the server workload would go back to 100%. It was as
if the SQL Server remembered the workstation and would try to do something
for it. (The next day this workstation did not have such an effect.) As I
said, Very Strange.
We also are on 32-bit using AWE memory, but only 12 GB. (We also only have
about 60 databases on the machine, nothing like your 1500+.) This weekend,
we are upgrading to 64-bit Windows 2003/SQL Server 2005. We will see if
this recurs, or whether it is an AWE flaw.
This thread from a Red Gate user suggests the problem is not free memory,
but free contiguous memory.
http://www.red-gate.com/MessageBoar...
b1ce133f1e6b1
So, if very large queries and cursor usage are responsible for fragmenting
memory (per something I read elsewhere) I don't know what to do except to
schedule periodic restarts of the the SQL Server.
RLF
<steve> wrote in message news:200722311948sgoodwin@.autotask.com...
> We are experiencing the same errors and same "lock out" behavior. We have
> the same version/patch level of SQL installed but are in a 32 bit env. The
> box has 32GB of memory. Nether the bPool or VAS show any pressure when the
> "out of system memory error" occurs. Since this error usually indicates a
> VAS problem we tried increasing VAS to 1GB w/the -g flag. It made no
> difference. PSS has been working on the case for 3 weeks. They dont seem
> to be finding any evidince of memory pressure either. When I last spole to
> the escalation engineer yesterday it seemed that they are going to
> recommend reducing the number of databases on the server. I asked for
> clarification as to whether we are hitting a 32 bit barrior, an instance
> limitation, or both. I am awaiting the answer. How many databases do you
> have on your server? We had between 1700 and 1900 (the number varies) at
> times when the error occured. We are now at 1500, and have not had the
> error in the 2 days since reducing the number of databases...
> EggHeadCafe.com - .NET Developer Portal of Choice
> http://www.eggheadcafe.com
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment