On 27-Jun-10 21:09:59, Chris G wrote:
On Sun, Jun 27, 2010 at 09:45:19PM +0100, Chris G wrote:
On Sun, Jun 27, 2010 at 09:25:28PM +0100, Chris G wrote:
On Sun, Jun 27, 2010 at 06:21:45PM +0100, Dennis Dryden wrote:
On Sun, Jun 27, 2010 at 5:51 PM, Chris G cl@isbd.net wrote:
I have an odd situation here, after a recent reboot my system seemed *very* slow indeed...
Try installing bootchart[1] it's perfect for this kind of problem and packaged up in Debian, Ubuntu and Fedora. If it was just the once maybe it was due to a file system check at boot?
A file system check takes almost an hour on my system, that's not what it was! :-)
I'll try bootchart, sounds useful.
... and it shows, amoung other things, that ureadahead sits for almost 20 seconds at startup in iowait with nothing else happening at all. Apart from that everything looks fairly normal.
... and dmesg shows the same delay:-
Jun 27 21:56:38 chris kernel: [ 2.842494] kjournald starting.Commit interval 5 seconds Jun 27 21:56:38 chris kernel: [ 2.842530] EXT3-fs: mounted filesystem with ordered data mode. Jun 27 21:56:38 chris kernel: [ 3.682882] type=1505 audit(1277672181.168:2): operation="profile_load" pid=443 name=/sbin/dhclien> Jun 27 21:56:38 chris kernel: [ 3.683497] type=1505 audit(1277672181.168:3): operation="profile_load" pid=443 name=/usr/lib/Netw> Jun 27 21:56:38 chris kernel: [ 3.683844] type=1505 audit(1277672181.178:4): operation="profile_load" pid=443 name=/usr/lib/conn> Jun 27 21:56:38 chris kernel: [ 3.718490] type=1505 audit(1277672181.208:5): operation="profile_load" pid=444 name=/usr/bin/evin> Jun 27 21:56:38 chris kernel: [ 3.728420] type=1505 audit(1277672181.218:6): operation="profile_load" pid=444 name=/usr/bin/evin> Jun 27 21:56:38 chris kernel: [ 3.734171] type=1505 audit(1277672181.228:7): operation="profile_load" pid=444 name=/usr/bin/evin> Jun 27 21:56:38 chris kernel: [ 3.755632] type=1505 audit(1277672181.248:8): operation="profile_load" pid=446 name=/usr/lib/cups> Jun 27 21:56:38 chris kernel: [ 3.756354] type=1505 audit(1277672181.248:9): operation="profile_load" pid=446 name=/usr/sbin/cup> Jun 27 21:56:38 chris kernel: [ 3.775301] type=1505 audit(1277672181.268:10): operation="profile_load" pid=447 name=/usr/sbin/my> Jun 27 21:56:38 chris kernel: [ 18.746128] udev: starting version 147 Jun 27 21:56:38 chris kernel: [ 18.763280] lp: driver loaded but no devices found Jun 27 21:56:38 chris kernel: [ 18.818800] Adding 19800072k swap on /dev/sda5. Priority:-1 extents:1 across:19800072k Jun 27 21:56:38 chris kernel: [ 18.851109] ip_tables: (C) 2000-2006 Netfilter Core Team Jun 27 21:56:38 chris kernel: [ 18.883661] r8169: eth0: link up Jun 27 21:56:38 chris kernel: [ 18.883667] r8169: eth0: link up Jun 27 21:56:38 chris kernel: [ 18.901562] alloc irq_desc for 22 on node 0 Jun 27 21:56:38 chris kernel: [ 18.901565] alloc kstat_irqs on node 0 Jun 27 21:56:38 chris kernel: [ 18.901574] HDA Intel 0000:00:1b.0: PCI INT A -> GSI 22 (level, low) -> IRQ 22 Jun 27 21:56:38 chris kernel: [ 18.901684] HDA Intel 0000:00:1b.0: setting latency timer to 64 Jun 27 21:56:38 chris kernel: [ 19.036644] input: HDA Digital PCBeep as /devices/pci0000:00/0000:00:1b.0/input/input5 Jun 27 21:56:38 chris kernel: [ 19.493924] EXT3 FS on sda1, internal journal Jun 27 21:56:38 chris kernel: [ 20.608497] kjournald starting. Commit interval 5 seconds
OK, it's a 15 second delay but still doing nothing for 15 seconds does seem a bit excessive.
-- Chris Green
Just a guess -- and not inspired by any of the above -- is it by any chance trying to NFSmount a filesystem from a different host, not getting a response, and waiting till it times out?
Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@manchester.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 28-Jun-10 Time: 00:26:31 ------------------------------ XFMail ------------------------------