I administer clamwin installed on a WinME system. I have been
unable to run a complete scan of the C-disk with any of the
cygwin-free versions -- 0.88.1, 0.88.2.1, 0.88.2.3. This is
the same for manual and scheduled scans.

When I attempt a full C-disk scan with any cygwin-free version,
everything seems to run fine for about 2.5 to 3.5 hours. Then...

- system response crawls to a stop
-- cursor movement response time slows to 5-10 seconds or longer
-- disk activity virtually stops
-- clock display fails to update (system time OK on reboot)

The system must be rebooted to recover. Even if I manage to kill
the clamwin process, the system is frozen, requiring reboot.

My guess is that the cygwin-free versions of clamwin use a buggy
WinME function that eventually exhausts memory and freeze system.

Version 0.88 works fine, so I have reverted to 0.88 for production.

Anyone else have this problem?

I have a similar problem. I do not know if it is the same thing, but here are my details:

I run ClamWin 0.88.2.3 on Windows XP. It is able to scan individual files and small folders just fine, but whenever I do a long scan (manual or scheduled), the system freezes after a couple of hours. The ClamWin scan remains stuck on a file (a different one every time, so I don't think it is a file problem), cursor movement ceases, and I cannot even bring up the Task Manager. I always end up needing to manually reboot.

Again, I don't know if this is the same problem or just one that shares certain characteristics. Still, if anyone has any ideas, please let me know.

It seams an huge memory leak, not easy traceable, also it spots only on winme that I don't have.
You can try to understand what's happening by monitoring the process and used resources with
process explorer from https://www.sysinternals.com/ https://www.sysinternals.com/

btw are you sure you don't want to drop winme? microsoft will drop the support in some months, also firefox will not be supported anymore on win9x arch

I'm working on pushing this user off of winme. In the meantime, I installed the
process explorer tools to take a deeper look at this problem. Not sure which will
happen first. Thanks for the pointers.

Using the process monitor (suggested by sherpya), I have verified that the original
problem occurs with no indication of a memory leak. Memory and CPU usage remain
stable, but the scan process slows to a crawl about 2.5 hours into the scan (as originally
described). It acts as if clamscan.exe starts executing a non-interruptible delay loop
that soaks up more and more CPU time. Does that make any sense to someone who
knows the code? If I stop the scan after cursor and clock sluggishness set in but before
the system freezes entirely, the OS continues to operate correctly after aborting the
scan. Now I'm thinking the problem may have nothing to do with winme. Note that
armand reported similar symptoms on an XP system. I will continue to try to isolate
the problem. Any thoughts on what to look for?

Just a thought : have you checked the energy saving settings of the system ? Especially the setting "Turn off hard disk" seems to be of interest here. You probably already checked this, but just to make sure...

Thanks for that interesting thought, budtse -- I had not checked this. As it turns
out, all power settings are set to "never", in particular "Turn off hard disk".