Gerald V. Livingston II <debuser@sysmatrix.net> [2002-11-03 13:34:01 -0600]: Please don't quote the entire message. Please trim your messages to reduce the clutter. > How about killing these? > root 31132 1 0 Oct31 ? 00:00:00 [usb-storage-0] > root 31133 1 0 Oct31 ? 00:00:00 [scsi_eh_2] You have hit a limit of my knowledge as this is linux kernel specific. I believe the brackets around the name mean that it has been completely swapped out to swap space and no longer is memory resident. The kernel can't tell us the real command line in this case as it has been stored on disk and would have to swap that process back in from disk in order to read the command line from it. Therefore the kernel reports only the name and lets us know it is partial by printing bracks around it. It does this instead of swapping everything back in just to get the command line because that would be inefficient and defeat the purpose of the long term scheduler, if just running ps brought them all back into memory. Or maybe it means it is a kernel process. Or perhaps either of those two. But neither of those are defunct. My previous note about killing defunct process' parents does not apply to these. But I am not sure what does either. > Yes, the device was unmounted and removed properly. Each time I start > the device back up it adds another entry and moves it's SCSI mount point > (/dev/sdd1 ---- /dev/sde1 --- /dev/sdf1). > > So far I've only found a reboot to clear the entries. I will guess that those are associated processes to a kernel module. Guessing that the mounting loaded the module but that nothing has unloaded the module. You might try 'lsmod' and 'rmmod'. But I am definitely guessing here. Bob
Attachment:
pgpQQRKS9cNVA.pgp
Description: PGP signature