[Sussex] New distro install (Mandriva 2006) problems - where too start?
Steve Dobson
steve at dobson.org
Sat Dec 10 06:47:55 UTC 2005
John
On Sat, Dec 10, 2005 at 05:26:26AM +0000, John D. wrote:
> Thomas Adam wrote:
> >So -- the computer is having to do lots od tasks. How is this unusual?
> >What you haven't said is what program you're using (Is
> >Man{driva,drake,whatever} still using DraxConf as an interface for this?
> >
> >
> Don't quite follow Thomas. With for example, the software manager and
> maybe firefox, or patience, swapping between windows is when the slow
> redrawing shows up - It doesn't have to be any particular app or
> combination of apps. Plus the slow screen redraw also shows at the
> taskbar sometimes. There seems to be no corrolation as to when the
> cursor becomes "jerky" other than that I've done "something" like click
> a button or scrolled a page.
What Thomas was saying, with out knowing your setup, that your computer
was busy. He assumed that there were processes running (maybe not
GUI ones) that were taking up CPU. Now you say that you are not doing
much - so the problem is probably else where - ie. not a busy system.
> >Desmond Armstrong wrote:
> >Please open a terminal window and then type su enter the root password
> >and then type 'top' and enter.
> >I have had the same problem but because of the extreme slowness I was
> >not able to establish the culprit.
> >Please report back the top cpu hoggers.
> >You may also go into the uninstall and remove 'kat'. This shows as
> >'kded' in top.
> >I suspect the problem is with xorg so I look forward to your findings.
> >With the machine I did I resorted to Mandriva 2005 and that runs very
> >well, but, I need to know what the problem is. Some information from
> >you and I shall retry Mandriva 2006.
> >Otherwise I am having great success with Mandriva2006
> >What is your processor speed and how much RAM do you have?
>
> The "top" command gave me this
I'll break the various output values from "top(1)"
> >top - 05:19:19 up 12:42, 3 users, load average: 0.29, 0.15, 0.05
This is just a overview line -
3 users logged in (your graphic shell and two command lines at a guess)
> >Tasks: 101 total, 1 running, 100 sleeping, 0 stopped, 0 zombie
A total of 101 task/jobs/processes running (inc some kernel threads)
> >Cpu(s): 1.0% us, 13.1% sy, 2.9% ni, 50.9% id, 31.4% wa, 0.4% hi, 0.3% si
Info about how the CPU is splitting up its time between the various
jobs:
us - Time spent running user processes (like your patience game or
firefox
sy - Time spent running system stuff (like getting data from disk
or the network - this generally should be lower the user
time - it is not - I am worried.
ni - Niced process time (a "nice"d process is one that has said
"I'm not important so only run me if you have nothing
better to do" - not much load there - see also idle
id - Idle time - the time the CPU had nothing to do, not even
low priority stuff.
wa - Waiting for I/O - the amount of time the CPU was waiting for
disk - This is way to high on your machine - something
to look into.
hi/si - Can't remember o the top of my head
> >Mem: 775504k total, 272108k used, 503396k free, 23876k buffers
> >Swap: 1469936k total, 101608k used, 1368328k free, 146008k cached
Summary of your memory usage physical and swap. Most of it is
free so no problems there
Below is a break down of the tasks running. The only one that is
registering any CPU time is X - given the summary above I would
suggest that this is a sick system.
> > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> > 3779 root 15 0 88784 7680 2076 S 7.9 1.0 5:59.61 X
> > 1 root 16 0 1564 76 56 S 0.0 0.0 0:38.24 init
> > 2 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0
> > 3 root 10 -5 0 0 0 S 0.0 0.0 0:00.05 events/0
> > 4 root 20 -5 0 0 0 S 0.0 0.0 0:00.01 khelper
> > 5 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 kthread
> > 7 root 20 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid
> > 84 root 10 -5 0 0 0 S 0.0 0.0 0:24.39 kblockd/0
> > 119 root 17 -5 0 0 0 S 0.0 0.0 0:00.00 aio/0
> > 118 root 15 0 0 0 0 S 0.0 0.0 5:59.67 kswapd0
> > 708 root 15 0 0 0 0 S 0.0 0.0 0:00.00 kseriod
> > 790 root 15 0 0 0 0 S 0.0 0.0 0:13.43 kjournald
> > 932 root 21 -4 1564 0 0 S 0.0 0.0 0:00.10 udevd
> > 1271 root 15 0 0 0 0 S 0.0 0.0 0:00.00 khubd
> > 1730 root 15 0 0 0 0 S 0.0 0.0 0:00.68 kjournald
> > 2734 root 16 0 1584 56 32 S 0.0 0.0 1:43.91 ifplugd
> > 2824 rpc 16 0 1688 0 0 S 0.0 0.0 0:00.00 portmap
> > 2848 root 16 0 1608 196 132 S 0.0 0.0 0:09.34 syslogd
> > 2860 root 16 0 2336 172 124 S 0.0 0.0 0:00.95 klogd
> > 2912 root 16 0 1552 0 0 S 0.0 0.0 0:00.00 acpid
> > 2958 root 20 0 1696 0 0 S 0.0 0.0 0:00.00 rpc.statd
> > 3052 root 16 0 5484 460 324 S 0.0 0.1 1:28.14 cupsd
> > 3462 xfs 16 0 3996 344 148 S 0.0 0.0 0:01.01 xfs
> > 3487 messageb 16 0 2224 0 0 S 0.0 0.0 0:00.01 dbus-daemon-1
> > 3512 root 17 0 4040 388 180 S 0.0 0.1 3:51.30 hald
> > 3605 root 16 0 1832 0 0 S 0.0 0.0 0:00.00 mandi
> > 3753 root 17 0 2612 196 152 S 0.0 0.0 0:00.00 kdm
> > 3754 root 16 0 2800 300 232 S 0.0 0.0 4:04.43 nifd
> > 3864 nobody 18 0 11424 96 68 S 0.0 0.0 0:00.11 mDNSResponder
> > 3890 daemon 16 0 1688 100 68 S 0.0 0.0 0:00.12 atd
> > 3933 root 18 0 4276 0 0 S 0.0 0.0 0:00.00 sshd
> > 3999 root 20 0 2148 0 0 S 0.0 0.0 0:00.00 xinetd
> > 4074 root 15 0 0 0 0 S 0.0 0.0 0:00.00 nfsd
> > 4075 root 15 0 0 0 0 S 0.0 0.0 0:00.00 nfsd
> > 4076 root 15 0 0 0 0 S 0.0 0.0 0:00.00 nfsd
> > 4077 root 15 0 0 0 0 S 0.0 0.0 0:00.00 nfsd
<rant>
This is why I don't like some of the standard installs. John is
a single computer user. So why does he need NFS running? I there
is another computer in his house I will probably be running Windows
so who has he got to share with?
</rant>
> Obviously I'm not familiar with that type of output, so don't know if
> that tells me/us/you if anything is particularly "sinful".
See above.
> My pc has a 2 gig pentium 4 chip and 768 megs of ram installed.
My laptop (that I'm working on now) is only a 1.13GHz with 128MB
and top gives me:
Cpu(s): 9.1% us, 1.3% sy, 0.0% ni, 87.6% id, 1.6% wa, 0.3% hi, 0.0% si
and I sometimes get summaries like this (the system is doing almost
nothing):
Cpu(s): 6.5% us, 0.6% sy, 0.0% ni, 92.5% id, 0.0% wa, 0.3% hi, 0.0% si
But this is a laptop and therefore has a SLOW laptop disk and I some
times see likes like this:
Cpu(s): 6.2% us, 6.9% sy, 11.8% ni, 43.6% id, 31.1% wa, 0.3% hi, 0.0% si
This is like yours - high "Wait or Disk" number compared to "user/system"
CPU times. But this type of summary is a rare event - so I do not
worry - and I expected it because I have very slow disks.
You might have just copied an odd bad line when your system at a very
brief peek at disk activity. This happens. top(1) is something that
you should what and see how the system changes over time. Of more help
in an e-mail is vmstat(8) which reports a number of samples of the system.
Try the following:-
$ vmstat 5 10
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 97424 7512 40 85020 2 4 14 13 3 463 7 1 91 0
0 0 97424 7396 40 85020 0 0 0 0 1001 190 1 0 99 0
0 0 97424 7388 40 85020 0 0 0 0 1001 195 1 0 99 0
0 0 97424 7388 40 85020 0 0 0 0 1001 198 1 0 99 0
0 0 97424 7380 40 85020 0 0 0 0 1162 374 3 1 96 0
0 0 97424 7404 40 85020 0 0 0 0 1028 503 1 1 98 0
0 0 97424 7388 40 85020 0 0 0 2 1020 437 2 0 98 0
0 0 97424 7388 40 85028 0 0 2 29 1026 460 2 0 98 0
0 0 97424 7388 40 85028 0 0 0 0 1018 465 2 0 98 0
0 0 97424 7404 40 85028 0 0 0 5 1160 355 3 1 96 0
> As I say, I've never experienced anything like this with any other
> distro - so I have to presume that it's something to do with this
> version of Mandriva. I've not selected/installed anything other than the
> default offerings and a few extra kde bits (the card games etc).
It may be that you have a problem with your disks or (more likly) that
your graphics system is not configured right - that you're not getting
full performance from the graphics hardware.
To check out the disk please report the output of:
$ dmesg | tail -20
This will give us the last 20 kernel messages - if a disk has problems
it should (?) show up there.
If the problem is graphics then the output logs from the Xserver is
what is needed. Have a look for a file like /var/log/Xorg.0.log or
/var/log/XFree86.0.log. It'll be big so send it to me personally
as an attachment. Also send me the output of "lspci -v"
Steve
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://mailman.lug.org.uk/pipermail/sussex/attachments/20051210/85ea27a4/attachment.pgp
More information about the Sussex
mailing list