Version: desconocido (using 4.2.3 (KDE 4.2.3), Kubuntu packages) Compiler: cc OS: Linux (x86_64) release 2.6.28-11-generic - Kubuntu Jaunty 64 bits - nvidia-drivers 177 - MD Turion(tm) 64 X2 Mobile Technology TL-52 - RAM 1 GB I'm really tired of having my system with a load average ~ 5 and SWAP usage when I have NO applications open (except KDE4, Kontact and some Konqueror, no more). This is the typicall output of "top" command when the system begins consumig resources (with no reason): top - 22:03:04 up 2:49, 1 user, load average: 4.79, 5.01, 4.22 Tasks: 145 total, 7 running, 138 sleeping, 0 stopped, 0 zombie Cpu(s): 30.5%us, 51.4%sy, 0.0%ni, 16.4%id, 1.0%wa, 0.3%hi, 0.3%si, 0.0%st Mem: 1012288k total, 999372k used, 12916k free, 74612k buffers Swap: 714884k total, 104260k used, 610624k free, 202148k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 3095 root 20 0 358m 142m 5184 S 4 14.4 7:31.08 Xorg 17304 root 20 0 171m 12m 1656 S 4 1.2 0:22.29 console-kit-dae 17397 haldaemo 20 0 37676 6116 3864 S 4 0.6 0:22.33 hald 11254 ibc 20 0 560m 79m 31m S 3 8.0 0:21.85 konqueror 17291 messageb 20 0 22224 1964 780 S 3 0.2 0:13.00 dbus-daemon 6296 ibc 20 0 312m 42m 23m S 2 4.3 0:21.31 kwin 7329 ibc 20 0 837m 62m 29m S 2 6.4 0:18.38 plasma 17398 root 20 0 15816 1256 1012 R 1 0.1 0:02.48 hald-runner 740 root 15 -5 0 0 0 S 0 0.0 0:05.58 kjournald 2644 syslog 20 0 12376 644 556 S 0 0.1 0:02.84 syslogd 10265 ibc 20 0 577m 31m 18m S 0 3.2 0:02.16 krunner 12034 ibc 20 0 233m 15m 11m S 0 1.5 0:00.60 klipper 13758 ibc 20 0 42572 2660 2068 S 0 0.3 0:00.03 ssh 17803 avahi 20 0 31888 1384 1096 S 0 0.1 0:00.61 avahi-daemon 17914 root 20 0 70988 2644 1876 S 0 0.3 0:01.21 NetworkManager I've no idea of why console-kit-dae, hald and dbus are consumig so much resources so often! It's really extrange for me the fact that "dstat" detects a high disk writting operations: ----total-cpu-usage---- -dsk/total- -net/total- ---paging-- ---system-- usr sys idl wai hiq siq| read writ| recv send| in out | int csw 14 16 68 2 0 0| 277k 184k| 0 0 | 11k 24k|1145 2044 32 49 17 2 0 1| 152k 1152k| 0 0 | 0 0 |1258 3477 29 51 19 0 0 0| 0 1088k| 0 0 | 0 0 |1195 3310 35 48 16 1 0 0| 0 1080k| 97B 81B| 0 0 |1334 3459 31 49 12 7 0 2|8664k 1016k|1856B 1120B| 0 0 |1389 3769 34 41 13 9 1 2| 19M 592k| 420B 132B| 0 0 |1740 3346 49 40 10 0 0 0| 0 952k| 420B 132B| 0 0 |1197 3029 45 48 7 1 0 0| 0 1024k| 420B 132B| 0 0 |1128 3200 ^[[C 50 37 11 0 0 0| 0 960k| 0 0 | 0 0 |1161 2868 43 46 11 0 1 0| 0 1056k| 0 0 | 0 0 |1206 3393 35 50 14 0 1 0| 0 1000k| 0 0 | 0 0 |1466 3791 29 55 15 1 0 0| 0 1072k| 0 206B| 0 0 |1281 3639 38 48 14 1 0 1| 0 832k|1572B 235B| 0 0 |1221 3202 31 52 16 1 0 1| 0 1152k| 523B 301B| 0 0 |1280 3646 34 48 16 1 0 0| 0 1136k| 420B 132B| 0 0 |1199 3396 32 52 16 1 0 0| 0 1056k| 420B 132B| 0 0 |1244 3634 33 55 11 1 0 1| 0 1176k| 420B 132B| 0 0 |1233 3623 30 53 16 1 0 1| 0 1152k| 420B 132B| 0 0 |1217 3443 32 51 15 1 0 0| 0 1184k| 420B 132B| 0 0 |1300 3569 31 50 17 1 0 0| 0 1088k| 420B 132B| 0 0 |1219 3646 Also, a "free -m" tellsme that SWAP is being used!: total used free shared buffers cached Mem: 988 974 13 0 85 207 -/+ buffers/cache: 681 306 Swap: 698 103 595 Perhaps this bug is not related to KDE, of course.
In order to show which processes are writting into the disk, I run: echo -n "1" >/proc/sys/vm/block_dump and this is what syslog says: May 31 22:12:34 ibc-laptop kernel: [10750.086373] kjournald(740): WRITE block 11502416 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086377] kjournald(740): WRITE block 11502424 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086380] kjournald(740): WRITE block 11502432 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086384] kjournald(740): WRITE block 11502440 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086387] kjournald(740): WRITE block 11502448 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086391] kjournald(740): WRITE block 11502456 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.086394] kjournald(740): WRITE block 11502464 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.087162] kjournald(740): WRITE block 11502472 on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.097671] sed(29572): dirtied inode 4415059 (sedN781FM) on tmpfs May 31 22:12:34 ibc-laptop kernel: [10750.120677] pidof(29578): dirtied inode 4415082 (exe) on proc May 31 22:12:34 ibc-laptop kernel: [10750.120744] pidof(29578): dirtied inode 4415085 (exe) on proc May 31 22:12:34 ibc-laptop kernel: [10750.120808] pidof(29578): dirtied inode 4415088 (exe) on proc May 31 22:12:34 ibc-laptop kernel: [10750.120870] pidof(29578): dirtied inode 4415091 (exe) on proc May 31 22:12:34 ibc-laptop kernel: [10750.148117] hal-acl-tool(29549): dirtied inode 4414876 (acl-list) on tmpfs May 31 22:12:34 ibc-laptop kernel: [10750.189596] pidof(29607): dirtied inode 4415231 (exe) on proc May 31 22:12:34 ibc-laptop kernel: [10750.194986] syslogd(2644): dirtied inode 287583 (auth.log) on sda1 May 31 22:12:34 ibc-laptop kernel: [10750.194995] syslogd(2644): dirtied inode 287583 (auth.log) on sda1
I suggest you first fix filesystem on your box. ext2fsck, ext3fsck, fsck & so on. You'll need to be root for this (the best is single user mode, esp. if you have so called rescue disc - use it now!) and unmount all fs's needed to be fixed. if you have classic (/, /tmp, /usr, /home, swap). Second, your system needs more RAM or more swap. You can organize swap on a usb key for temporary matters - to see which size fits you the best. Cheers, Jože
You could also try to update to the latest nVidia drivers as older versions has known bugs and memory leaks.
Thanks to all. So it seems it's not a KDE related bug. I'll check the disk of course. @Dario: Nvidia newer version (180) has a critial bug: https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-180/+bug/363672
The bug report page mentions "180.44". However the most recent version is "180.60" (at least on http://www.nvidia.com/object/unix.html). But they have a different way of installing, so it's ok if you are not willing to try them. Regards!
Yes, I prefer having the Ubuntu install way, even with privative drivers. However I'll inspect if they have fixed the bug I mentioned. thanks.