Re: Estimate Timeout Issue - Dump runs fine

[Tom Brown]

OK thanks - I have increased the etimeout to 2400 seconds and also 
changed the udp timeout within checkpoint to also be 2400 seconds so 
i'll see how the run goes tonight


everything was fine today - no estimate timeout

thanks for the pointer

Estimate Timeout Issue - Dump runs fine

[Tom Brown]

Hi

Server is 2.4.5 and client is now 2.4.5p1 both on CentOS

I use Amanda and have done for years with no issues setting up etc - I 
can pretty much set up with my eyes closed now!! Amanda rocks...


But i'm getting a slightly strange error with a large partition. The 
partition in question is around 900gig in size although only a few 
hundred meg are currently used. When the estimate runs it returns


FAILURE AND STRANGE DUMP SUMMARY:
  planner: ERROR Estimate timeout from servername

Thing is though the actual dump of this filesystem runs fine - I have 
increased my eTimeout to 20mins but this still occurs - Any ideas on 
this one?


thanks

Re: Estimate Timeout Issue - Dump runs fine

[Joshua Baker-LePain]

On Wed, 2 Nov 2005 at 11:32am, Tom Brown wrote

But i'm getting a slightly strange error with a large partition. The 
partition in question is around 900gig in size although only a few hundred 
meg are currently used. When the estimate runs it returns


FAILURE AND STRANGE DUMP SUMMARY:
 planner: ERROR Estimate timeout from servername

Thing is though the actual dump of this filesystem runs fine - I have 
increased my eTimeout to 20mins but this still occurs - Any ideas on this 
one?


Look in /tmp/amanda/sendsize*debug and/or amandad*debug to see how long 
the estimate is actually taking.  Also, what do your iptables rules look 
like on the server?


--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University

Re: Estimate Timeout Issue - Dump runs fine

[Tom Brown]

Look in /tmp/amanda/sendsize*debug and/or amandad*debug to see how long 
the estimate is actually taking.  Also, what do your iptables rules look 
like on the server?


thanks - iptables are not being used, local firewall is off

sendsize degug is below and looks OK

# more /tmp/amanda/sendsize.20051102003001.debug
sendsize: debug 1 pid 12320 ruid 11 euid 11: start at Wed Nov  2 
00:30:01 2005

sendsize: version 2.4.5p1
sendsize[12322]: time 0.002: calculating for amname '/dev/sda2', dirname 
'/', spindle -1

sendsize[12322]: time 0.002: getting size via dump for /dev/sda2 level 0
sendsize[12322]: time 0.002: calculating for device '/dev/sda2' with 'ext3'
sendsize[12322]: time 0.002: running /sbin/dump 0Ssf 1048576 - /dev/sda2
sendsize[12322]: time 0.003: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12320]: time 0.003: waiting for any estimate child: 1 running
sendsize[12322]: time 21.884: 1447269376
sendsize[12322]: time 21.885: .
sendsize[12322]: estimate time for /dev/sda2 level 0: 21.882
sendsize[12322]: estimate size for /dev/sda2 level 0: 1413349 KB
sendsize[12322]: time 21.885: asking killpgrp to terminate
sendsize[12322]: time 22.886: getting size via dump for /dev/sda2 level 1
sendsize[12322]: time 22.887: calculating for device '/dev/sda2' with 'ext3'
sendsize[12322]: time 22.887: running /sbin/dump 1Ssf 1048576 - /dev/sda2
sendsize[12322]: time 22.888: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12322]: time 195.606: 4647936
sendsize[12322]: time 195.606: .
sendsize[12322]: estimate time for /dev/sda2 level 1: 172.718
sendsize[12322]: estimate size for /dev/sda2 level 1: 4539 KB
sendsize[12322]: time 195.606: asking killpgrp to terminate
sendsize[12322]: time 196.608: done with amname '/dev/sda2', dirname 
'/', spindle -1

sendsize[12320]: time 196.608: child 12322 terminated normally
sendsize[12334]: time 196.609: calculating for amname '/dev/sda1', 
dirname '/boot', spindle -1

sendsize[12334]: time 196.609: getting size via dump for /dev/sda1 level 0
sendsize[12334]: time 196.609: calculating for device '/dev/sda1' with 
'ext3'

sendsize[12334]: time 196.609: running /sbin/dump 0Ssf 1048576 - /dev/sda1
sendsize[12320]: time 196.609: waiting for any estimate child: 1 running
sendsize[12334]: time 196.610: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12334]: time 197.239: 5737472
sendsize[12334]: time 197.239: .
sendsize[12334]: estimate time for /dev/sda1 level 0: 0.630
sendsize[12334]: estimate size for /dev/sda1 level 0: 5603 KB
sendsize[12334]: time 197.239: asking killpgrp to terminate
sendsize[12334]: time 198.242: getting size via dump for /dev/sda1 level 1
sendsize[12334]: time 198.243: calculating for device '/dev/sda1' with 
'ext3'

sendsize[12334]: time 198.243: running /sbin/dump 1Ssf 1048576 - /dev/sda1
sendsize[12334]: time 198.243: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12334]: time 198.684: 27648
sendsize[12334]: time 198.684: .
sendsize[12334]: estimate time for /dev/sda1 level 1: 0.441
sendsize[12334]: estimate size for /dev/sda1 level 1: 27 KB
sendsize[12334]: time 198.684: asking killpgrp to terminate
sendsize[12334]: time 199.687: done with amname '/dev/sda1', dirname 
'/boot', spindle -1

sendsize[12320]: time 199.687: child 12334 terminated normally
sendsize[12339]: time 199.687: calculating for amname '/dev/sda5', 
dirname '/export/disk1', spindle -1

sendsize[12339]: time 199.688: getting size via dump for /dev/sda5 level 0
sendsize[12320]: time 199.688: waiting for any estimate child: 1 running
sendsize[12339]: time 199.688: calculating for device '/dev/sda5' with 
'ext3'

sendsize[12339]: time 199.688: running /sbin/dump 0Ssf 1048576 - /dev/sda5
sendsize[12339]: time 199.689: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12339]: time 545.606: 88973312
sendsize[12339]: time 545.617: .
sendsize[12339]: estimate time for /dev/sda5 level 0: 345.928
sendsize[12339]: estimate size for /dev/sda5 level 0: 86888 KB
sendsize[12339]: time 545.617: asking killpgrp to terminate
sendsize[12339]: time 546.619: getting size via dump for /dev/sda5 level 1
sendsize[12339]: time 546.646: calculating for device '/dev/sda5' with 
'ext3'

sendsize[12339]: time 546.646: running /sbin/dump 1Ssf 1048576 - /dev/sda5
sendsize[12339]: time 546.647: running /opt/amanda-2.4.5p1/libexec/killpgrp
sendsize[12339]: time 2182.684: 25811968
sendsize[12339]: time 2182.696: .
sendsize[12339]: estimate time for /dev/sda5 level 1: 1636.054
sendsize[12339]: estimate size for /dev/sda5 level 1: 25207 KB
sendsize[12339]: time 2182.701: asking killpgrp to terminate
sendsize[12339]: time 2183.703: done with amname '/dev/sda5', dirname 
'/export/disk1', spindle -1

sendsize[12320]: time 2183.704: child 12339 terminated normally
sendsize: time 2183.704: pid 12320 finish time Wed Nov  2 01:06:24 2005

one of my amanda.debugs does have this at the bottom of it

amandad: time 2193.716: dgram_recv: timeout after 10 seconds
amandad: time 

Re: Estimate Timeout Issue - Dump runs fine

[Joshua Baker-LePain]

On Wed, 2 Nov 2005 at 2:31pm, Tom Brown wrote

Look in /tmp/amanda/sendsize*debug and/or amandad*debug to see how long the 
estimate is actually taking.  Also, what do your iptables rules look like 
on the server?


thanks - iptables are not being used, local firewall is off



one of my amanda.debugs does have this at the bottom of it

amandad: time 2193.716: dgram_recv: timeout after 10 seconds
amandad: time 2193.716: waiting for ack: timeout, retrying
amandad: time 2203.716: dgram_recv: timeout after 10 seconds
amandad: time 2203.716: waiting for ack: timeout, retrying
amandad: time 2213.717: dgram_recv: timeout after 10 seconds
amandad: time 2213.717: waiting for ack: timeout, retrying
amandad: time 2223.717: dgram_recv: timeout after 10 seconds
amandad: time 2223.717: waiting for ack: timeout, retrying
amandad: time 2233.718: dgram_recv: timeout after 10 seconds
amandad: time 2233.718: waiting for ack: timeout, giving up!
amandad: time 2233.718: pid 12319 finish time Wed Nov  2 01:07:14 2005

is that time figure a time in seconds ?


Yep.  So you can just increase etimeout and/or figure out why /sbin/dump 
1Ssf 1048576 - /dev/sda5 is taking so long.


--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University

Re: Estimate Timeout Issue - Dump runs fine

[Tom Brown]

Yep.  So you can just increase etimeout and/or figure out why 
/sbin/dump 1Ssf 1048576 - /dev/sda5 is taking so long.


OK thanks - I have increased the etimeout to 2400 seconds and also 
changed the udp timeout within checkpoint to also be 2400 seconds so 
i'll see how the run goes tonight


thanks