Hi.

Tracker is killing my machine (eating too much CPU and memory, which was already reported in #467270), so I pause it (from the applet).

But still, even when paused, top shows it as using more than 90% of my CPU... what's wrong ?

The log doesn't show any activity : last message is :
27 mar 2009, 08:15:43: Tracker: New DBus request, not pausing indexer, already in paused state
27 mar 2009, 08:15:43: Tracker: ---> [35] Success, no error given

What the heck is it doing when paused : polling resources still ?

I issued a strace -p on a running trackerd and saw :
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679201}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679302}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679403}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679503}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679604}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679704}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679810}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 679909}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 680010}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 680109}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1238139027, 680209}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}, {fd=9, events=POLLIN}], 5, 0) = 0 (Timeout)

nice flood...
...

Why is it polling constantly some resource instead of say every few second maybe ?... dunno :(

Thanks in advance.

Best regards,

Hi.

This happens quite systematically... any hope of a solution, workaround, or at least any clue that I'm not alone ?

Thanks in advance.

Best regards,

It seems that the default options are fare too aggressive, and starting over with nicer ones (see #525107) helps workaround the issue.

Hope this helps,

Hello,

I also suffer from this bug.

I modified my config some days ago as indicated in #525107 so that
tracker is much more bearable when indexing things. And it is quite
unintrusive now for most workloads.

However yesterday I paused tracker because I had to download some rather
large files, and it was constantly reindexing them, which impacted my
laptop quite a bit even with the above changes. So tracker is paused.

The laptop was put to sleep for the night, awakened this morning, and
trackerd is at 100%, while still paused.

In the log, I get this line repeated ad nauseam :

27 mai 2009, 09:27:17: Tracker: We are paused, sending nothing to the
index until we are unpaused

I can understand that tracker is eating many cpu cycles when indexing,
and that the configuration needs tweaking to make it less intrusive. But
when paused ???

Thanks,

Hi,

I cannot reproduce this issue with tracker 0.8.12-1 from unstable.
Could you please upgrade to the latest 0.8 version from unstable/testing and try
to reproduce this particular problem.

Thanks,
Michael