Re: Apps swapped out without swap or what?



On Aug 26, 9:17 pm, linuxnewbie1234 <linuxnewbie1...@xxxxxxxxxxx>
wrote:
Strange problem I can't understand...

I have a new dualcore computer. Notice the dualcore.

There was trackerd running today, indexing the HD, and the computer was
behaving slowly.

I had already niced trackerd to 19, ioniced it to idle class (CFQ
scheduler!), and I HAVE NO SWAP PARTITION in this computer!

Notwithstanding all this, when I would return to the shell after a
minute of inactivity or so, the shell had problems responding, I had to
wait seconds like if it was swapped out.

In some cases I noticed Firefox was even worse.

How could that happen??

There are a number of possible causes. You certainly haven't given us
enough details to provide a diagnosis of your specific problem, but
off the top of my head, you
a) might be suffering /because/ you have no swap. Without swap,
processes must wait until real memory becomes available, in order to
load pages or otherwise allocate memory. With swap, real memory can be
made available quickly by moving blocks that haven't been referenced
in a while to the swap area, thus freeing up the memory pages for new
use. Without swap, no swap area, no movement of older pages, and new
page requests have to wait until the process that owns the memory
releases it.
b) might be suffering from some TCP/IP-related problem, like a domain-
name resolution issue. Non-responsive TCP/IP connections take seconds
to time out, and if the shell, for instance, is trying to resolve the
hostname (in order to put it out on your command prompt), a non-
responsive DNS could cause the shell to hang until the TCP connection
times out
c) might be suffering from conflicting I/O, where there are many I/O
requests queued for one device or one area on a device, and those I/O
commands are delaying the I/O commands that the program loader uses to
retrieve program code segments for execution. If your shell needs a
piece of ELF module that hasn't been loaded yet, you may wait until
the other I/O is complete before the page is retrieved. (This should
be a subsecond wait, but under heavy load, or with large I/O, you
might be in "D" status for a while)
d) You hit the ^S by accident, and suspended output from the shell.
You subsequently hit ^Q (again by accident) to resume output


Ok to say the truth I also have an encrypted HD and the dm-crypt (or
whatever similar) process is root and nice is -5 and trackerd was
sucking data from those processes however I think it shouldn't behave
like hanged for seconds like if the applications were swapped out.


.



Relevant Pages

  • Re: Server REALLY slow after console messages
    ... on that server has used it all, including all your swap. ... RAM, when memory was relatively a lot slower and way more expensive, ... Boost the priority of your shell. ...
    (Debian-User)
  • Re: controlling mmap()d vs read/write() pages
    ... If not then they are not memory attacks ... I completely concede that you can DOS the system with I/O if that is ... My point is that is not a memory problem but a disk I/O problem which is ... It can be done trivially without performing any IO or swap, ...
    (Linux-Kernel)
  • Re: running Linux with no swap space (but lots of RAM)
    ... | capable of tolerating paging I/O, not that the system is paging to ... Please explain what you mean by "tolerating paging I/O". ... to swap anything. ... The amount of memory allocated to ...
    (comp.os.linux.development.system)
  • Re: running Linux with no swap space (but lots of RAM)
    ... You'd be better off having 8GB of RAM and 4GB of swap. ... it takes I/O time to write it as well as read it ... other usage of memory is causing this, ...
    (comp.os.linux.development.system)
  • Re: Heaps and Foreigners
    ... >> can ask for the memory and get a pointer to it. ... cause unfathomable deaths when swap space gets low or is exhausted. ... It's probably because there is some reserve memory to print out errors ... of RAM and disk space) and virtual swap (which is just a number managed ...
    (comp.lang.lisp)