Re: how to interpret poolmon output, 'Proc' tag.



three bits of information.
1a) You might get temporary remedy by increasing the nonpaged pool
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager
\Memory Management
by changing the value of NonPagedPoolSize. But as far as I understand,
the minimum nonpaged pool size is actually 128MB. When you increase
this parameter, the maximum reserved area for nonpaged pool will
increase a bit too. How the numbers are related, I do not know. It is
possible to monitor the sizes of reserved nonpaged and paged pool
areas with Process Explorer, when debug symbols are installed, under
System Information menu item.
1b) In my case there is also leak under pagedpool, under pool tag
"Toke", which means "Token objects". So You may need to increase the
maximum of paged pool too, in case it gets depleted too. I wont
describe the registry changes here much longer unless You ask for it.
There are other issues that need to be taken account when changing
these parameters. For example "System Page Table Entries" may become
low, but it is possible to monitor it with performance.msc and if
neccessary it is possible to find new reasonable configuration.
2. What was recently changed in my computer, is the following: 1)
installed newest ATI video card driver, 2) installed RATT and
kernrate, 3) installed some usual MS hotfixes. But my previous last
reboot was a month ago, so there might have been more changes during
that month; changes that I am not aware of right now. Anyway, before
my previous last reboot this problem did not occur, its a very recent
development which has been manifesting since reboot on last weekend.
3) PID-s may seometimes be large. When they do, they start being
systematically large, not only few of them. I have seen it too. It
seems to occur seldom, but randomly, even when no other bad things are
manifesting. Still there may be some relation, because right now my
PID numbers are in 100'000-s too.


ama...@xxxxxxxxx wrote:
On 21 abr, 20:52, levitation <roland.pihla...@xxxxxxxxx> wrote:
One more thing. It might help you if you disable any scheduled or
otherwise reoccuring tasks.
The Proc tag leaks at least in my case always when a new process
starts (it does not restore itself after closing the process). For
usual daily activities with a few programs, its so small that you wont
notice. But when some process starts and stops repeatedly in a
scheduled manner, this leak accumulates faster.


I'm having the same problem, we run 4 processes through our
application task scheduler every 10 second and i see a kernel memory
leak caused by the Proc tag pool, forcing to reboot the server every 2
days.
Another weird thing is the process identifier PID in the task manager
for new processes, the number is around 250.000 and growing
.

Relevant Pages

  • Re: how to interpret poolmon output, Proc tag.
    ... I indeed also have a leak associated to the 'Toke' tag in the paged pool. ... and reboot my computer regularly. ... the minimum nonpaged pool size is actually 128MB. ...
    (microsoft.public.windowsxp.general)
  • Re: Pagefile use increase when repeatedly executing programs
    ... Using Performance Monitor to Identify a Pool Leak ... as my pagefile use steadily increases, ... scanner, updated my computer bios, defragmented memory and pagefile, ... You could confirm that by doing a perfmon on the Nonpaged Pool objects ...
    (microsoft.public.windowsxp.general)
  • Re: Event ID 2017
    ... Adding a switch should not make any difference to this, I don't think that would be the problem, but maybe when you added the switch you changed another setting by mistake. ... You can also monitor the Server Object Pool Counters to see if anything there looks amiss, specifically the Pool Nonpaged Failures and the Pool Paged Failures, if you get values of 1 or more on these counters you probably don't have enough RAM in the machine. ... "The server was unable to allocate from the system nonpaged pool because the server reached the configured limit for nonpaged pool allocations." ...
    (microsoft.public.windowsxp.help_and_support)
  • Re: Exchange HTTP Virtual Server Instance 100 :The IsAlive check f
    ... The Free System Page Table Entries never drop below 8400. ... -Warning" when the Pool Paged Bytes counter exceeds 200 MB ... -Critical" when the Pool Paged Bytes counter exceeds 220 MB ... Nonpaged Pool Bytes, but it did not cross the "critical" threshold. ...
    (microsoft.public.windows.server.clustering)
  • Re: Exchange HTTP Virtual Server Instance 100 :The IsAlive check f
    ... The Free System Page Table Entries never drop below 8400. ... -Warning" when the Pool Paged Bytes counter exceeds 200 MB ... Nonpaged Pool Bytes, but it did not cross the "critical" threshold. ... cluster server might think the HTTP virtual server is "offline"? ...
    (microsoft.public.windows.server.clustering)