Memory leak??

[DR1]

After installing the new version and letting the computer sit undisturbed for a couple days I have noticed that memory usage is slowly creeping up for cpf.exe. What started out at 15,080 is now at 17,746. Anyone else seeing this?

[kail]

No. At the moment my W2k system has been up for 2 days & 3.5 hours. Currently, CPF.EXE is 3,352KB in memory (with 56,208KB in total) & CMDAGENT.EXE is 8,912KB in memory (38,060KB in total).

Are your sizes “in memory” or are they the total Virtual Size?

My “in memory” sizes mean actually in memory, whilst the rest is swapped.

[DR1]

It is actual “Memory Utilization”. cmdagent is 8,220 K actual.

[DR1]

This doesn’t look good…(see attachment)

[attachment deleted by admin]

[kail]

I’m not convinced it is a memory leak, as such.

The reason why I suspect this is as follows (something I’ve not reported yet). My system is W2k & if I reboot my system without manually shutting down CPF first, then I will loose all the entries in the CPF log since the last time it was closed (not on a reboot). To me, this implies that CPF is not committing the log entries until it is closed. But, on my system… despite giving all apps 60 seconds grace on shutdown… I still loose all the log entries on reboot, if I don’t close CPF manually first.

So… if you’re still following my nutty logic, then it seems that CPF is storing things in memory. This alone could explain memory growth over time.

I think this thread needs the man… the egemen that is. 8)

[egemen]

@kail post:5

I’m not convinced it is a memory leak, as such.

The reason why I suspect this is as follows (something I’ve not reported yet). My system is W2k & if I reboot my system without manually shutting down CPF first, then I will loose all the entries in the CPF log since the last time it was closed (not on a reboot). To me, this implies that CPF is not committing the log entries until it is closed. But, on my system… despite giving all apps 60 seconds grace on shutdown… I still loose all the log entries on reboot, if I don’t close CPF manually first.

So… if you’re still following my nutty logic, then it seems that CPF is storing things in memory. This alone could explain memory growth over time.

I think this thread needs the man… the egemen that is. 8)

Yes it may be the logs if there are a huge number of them and max log size is set to something more than 5 MB. But it should not affect cmdagent.exe at all.

Egemen

[kail]

@egemen post:6

Yes it may be the logs if there are a huge number of them and max log size is set to something more than 5 MB. But it should not affect cmdagent.exe at all.

Thanks for your quick response egemen. I can also understand why cpf.exe memory allocation might grow over time (without use?). But, in DRs previous posts, DR said that the memory allocation of the cmdagent.exe appeared to be growing… without being used (not used for a couple of days). Is this possible?

[egemen]

@kail post:7

Thanks for your quick response egemen. I can also understand why cpf.exe memory allocation might grow over time (without use?). But, in DRs previous posts, DR said that the memory allocation of the cmdagent.exe appeared to be growing… without being used (not used for a couple of days). Is this possible?

It may be possible if too many different applications + components connect to the internet. But not tens of MBs of growth ofcourse.

[system]

@kail post:7

Thanks for your quick response egemen. I can also understand why cpf.exe memory allocation might grow over time (without use?). But, in DRs previous posts, DR said that the memory allocation of the cmdagent.exe appeared to be growing… without being used (not used for a couple of days). Is this possible?

Hmmm. I did test this myself and run about 9200 standard usage, the only time it went up was if something was accessing the internet, if I was on it, or had many browser windows open, 42% CPF, and slight fluctuation % here and there. When I would close the browser windows, etc…it would drop to normal. When my auto updater on my anti-v was going and left the ask window up, this shot up as well to around 38% . Total mem usage between cpf etc…well, look here…any thoughts?

Cheers,

Paul

[attachment deleted by admin]

[DR1]

@kail post:7

Thanks for your quick response egemen. I can also understand why cpf.exe memory allocation might grow over time (without use?). But, in DRs previous posts, DR said that the memory allocation of the cmdagent.exe appeared to be growing… without being used (not used for a couple of days). Is this possible?

Actually, if you refer to my original post you will see that it is cpf.exe that is growing in size. The cmdagent.exe is remaining fairly consistant.

[DR1]

@egemen post:6

Yes it may be the logs if there are a huge number of them and max log size is set to something more than 5 MB. But it should not affect cmdagent.exe at all.

Egemen

There are no logs and the size is set to the default 5MB.

[kail]

Since I had egemen on-line… I did my best, in your absence, to advance your issue. Sorry, if I got off track with the cmdagent. Let’s see what egemen says when he sees that your log file was set to 5MB & EMPTY.

[DR1]

@kail post:12

Since I had egemen on-line… I did my best, in your absence, to advance your issue. Sorry, if I got off track with the cmdagent. Let’s see what egemen says when he sees that your log file was set to 5MB & EMPTY.

Thanks for your input on this, it is appreciated.

I rebooted last night after the cpf.exe went over 29,000 K. On reboot it was around 15,500 (I don’t remember the precise number), anyway, I will see where it is when I get home tonight.

thanks again,
DR

[egemen]

@DR post:13

Thanks for your input on this, it is appreciated.

I rebooted last night after the cpf.exe went over 29,000 K. On reboot it was around 15,500 (I don’t remember the precise number), anyway, I will see where it is when I get home tonight.

thanks again,
DR

Hmmm. If I recall correct, you are using ATI Catalyst(CLI.EXE). Can you list the DLLs loaded into CPF memory by using process explorer from www.sysinternals.com. I have seen this type of behavior reported previously before. It may be a case that this ATI control center places a global hook which maps a DLL file into CPF"s memory, which causes this behavior.

If you can list all the DLLs in CPF with process explorer, we can see what is going on.

Thx,
Egemen

[kail]

I think you could be right Egemen… DR had 3 CLI.EXEs in a previous example he posted.

[system]

@kail post:15

I think you could be right Egemen… DR had 3 CLI.EXEs in a previous example he posted.

Hi, kail,

I would assume then that DR an myself don’t have same existing issues? The only reason I posted that was to compare and perhaps see if when I had anything accessing the net the usage would shoot up considerably with browser slowdown. I don’t have 3 CLI’s so I think safe to assume these are not the same issues so I will throw my issue out and put it elsewhere…

Cheers,

Paul

[kail]

@comicfan2000 post:16

I would assume then that DR an myself don’t have same existing issues? The only reason I posted that was to compare and perhaps see if when I had anything accessing the net the usage would shoot up considerably with browser slowdown. I don’t have 3 CLI’s so I think safe to assume these are not the same issues so I will throw my issue out and put it elsewhere…

Yes, I think this is different. DR was saying that CPFs memory allocation was growing even though the PC in question had not been used.

[system]

@kail post:17

Yes, I think this is different. DR was saying that CPFs memory allocation was growing even though the PC in question had not been used.

Alrighttyy then, thank you.

Cheers,

Paul

[DR1]

@egemen post:14

Hmmm. If I recall correct, you are using ATI Catalyst(CLI.EXE). Can you list the DLLs loaded into CPF memory by using process explorer from www.sysinternals.com. I have seen this type of behavior reported previously before. It may be a case that this ATI control center places a global hook which maps a DLL file into CPF"s memory, which causes this behavior.

If you can list all the DLLs in CPF with process explorer, we can see what is going on.

Thx,
Egemen

I’ll take a look at that tonight.

Thanks
DR

[DR1]

@egemen post:14

Hmmm. If I recall correct, you are using ATI Catalyst(CLI.EXE). Can you list the DLLs loaded into CPF memory by using process explorer from www.sysinternals.com. I have seen this type of behavior reported previously before. It may be a case that this ATI control center places a global hook which maps a DLL file into CPF"s memory, which causes this behavior.

If you can list all the DLLs in CPF with process explorer, we can see what is going on.

Thx,
Egemen

I found one dll in ATI Catalyst CLI.exe that was hooked into Comodo. It was named appguard.dll (see screenshot)

Memory seems to be stable most of the time now for at around 15,400 for cpf.exe. The only time I am seeing this go higher is after I open the firewall and do anything, like modifying a rule or something as simple as checking status. After that I see memory usage start creeping up with each access of the firewall. It’s small increases but the thing that is unusual is that it stays there and doesn’t go back down unless I reboot. Right now it is running at 18,304.

[attachment deleted by admin]