extreme file growth fills disk re: tracker

Asked by Steve Reith

On 12262007 I found that that my "gutsy" had the entire disk partition filled with unknown data. I found this at login - showed a pop up indicator that there is no more disk space available. My gutsy partition is ext3 and 40 GB out of approx 300GB for other linux releases. There is only a / filesystem.

<du -a |sort -n> found that the evolution directory " /home/steve/.cache/tracker" looked like so -

root@ub-smr:/home/steve/.cache/tracker# ls -al
total 36208848
drwxr-xr-x 2 steve steve 4096 2007-12-26 23:46 .
drwxr-xr-x 3 steve steve 4096 2007-10-19 21:16 ..
-rw------- 1 steve steve 1744896 2007-12-21 19:46 email-contents.db
-rw------- 1 steve steve 3427848 2007-12-26 23:29 email-index.db
-rw------- 1 steve steve 724992 2007-12-21 20:23 email-meta.db
-rw------- 1 steve steve 9281536 2007-12-26 23:45 file-contents.db
-rw------- 1 steve steve 2147483647 2007-12-26 23:29 file-index.db
-rw------- 1 steve steve 16548560 2007-12-21 22:33 file-index.tmp.1
-rw------- 1 steve steve 34861209600 2007-12-26 23:35 file-meta.db
-rw------- 1 steve steve 1179784 2007-12-26 23:29 file-update-index.db

Needless to say evolution did not work - or for that matter any other application - I was lucky the unix command did

My quick fix was to empty -rw------- 1 steve steve 34861209600 2007-12-26 23:35 file-meta.db

like so : root@ub-smr:/home/steve/.cache/tracker# > file-meta.db

same with - root@ub-smr:/home/steve/.cache/tracker# > file-index.db

The disk is now back to 40 GB and evolution operates without any apparent errors.

root@ub-smr:/home/steve/.cache/tracker# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sdb2 46251812 10764724 34077388 25% /
varrun 647508 96 647412 1% /var/run
varlock 647508 0 647508 0% /var/lock
udev 647508 96 647412 1% /dev
devshm 647508 0 647508 0% /dev/shm
lrm 647508 34696 612812 6% /lib/modules/2.6.22-14-generic/volatile

That doesn't fix the underlying issue and I have no idea what evolution is doing.

This is a FYI and I would appreciate any and all feedback

regards,

Steve Reith

Question information

Language:
English Edit question
Status:
Solved
For:
Ubuntu Edit question
Assignee:
No assignee Edit question
Solved by:
Andrea Corbellini
Solved:
Last query:
Last reply:
Revision history for this message
C de-Avillez (hggdh2) said :
#1

This does not seem to be Evo, but tracker (a file indexing package). Please check if you have tracker installed, and check (if it is indeed installed) what are the options you set on it.

Revision history for this message
Steve Reith (steve-reith) said :
#2

Thanks for your quick response - the version of tracker I have is very
basic and I don't know how to configure any options - no menus

On Fri, 2007-12-28 at 00:44 +0000, hggdh wrote:
> Your question #20822 on Ubuntu changed:
> https://answers.edge.launchpad.net/ubuntu/+question/20822
>
> Status: Open => Needs information
>
> hggdh requested for more information:
> This does not seem to be Evo, but tracker (a file indexing package).
> Please check if you have tracker installed, and check (if it is indeed
> installed) what are the options you set on it.
>

Revision history for this message
marcobra (Marco Braida) (marcobra) said :
#3

Please take a look into your user tracker config file, and modify it:

gedit $HOME/.config/tracker/tracker.cfg

Hope this helps

Revision history for this message
C de-Avillez (hggdh2) said :
#4

Still... on a 40G total available disk space, tracker indexed 34G of Evolution emails... that's weird. What version of tracker are you running? Or are you indexing other filesystems?

Revision history for this message
Steve Reith (steve-reith) said :
#5

I think I found the issue - or at least another problem - there is a
memory leak with the daemon - currently on my machine with tracker app
not being used it is 1GB memory and climbing

<from top>

23488 steve 34 19 1154m 1.0g 1384 S 38.6 84.7 375:09.93 trackerd-

also vmstat shows severe thrashing from si so paging to disk

procs -----------memory---------- ---swap-- -----io---- -system--
----cpu----
 r b swpd free buff cache si so bi bo in cs us sy
id wa
 0 2 755324 26412 848 27788 13 14 58 103 174 355 69 22
8 1
 1 2 755400 28080 840 27832 240 1000 240 1024 356 307 72 13
0 15
 0 2 755296 27980 856 28032 348 0 572 104 304 544 37 16
0 48
 2 1 755032 25188 920 30316 536 0 2884 0 273 528 76 7
0 17
 1 1 755032 26384 916 30324 540 304 540 304 361 433 86 14
0 0
 0 1 755032 25996 916 30316 492 0 492 0 217 559 42 20
0 38
 2 1 755780 27400 920 30176 184 2172 184 2280 352 329 75 10
0 15
 0 1 755808 27804 924 30040 176 696 180 696 377 302 29 7
0 64
 1 1 755692 27692 924 30084 400 0 400 0 229 565 68 16
0 16
 1 1 755404 27036 924 30040 664 0 664 0 241 396 89 11
0 0

the tracker site speaks of 063 and 064 memory leak fixes - but I think
they still have the problem.

regards,

Steve Reith

On Fri, 2007-12-28 at 12:49 +0000, hggdh wrote:
> Your question #20822 on Ubuntu changed:
> https://answers.edge.launchpad.net/ubuntu/+question/20822
>
> hggdh requested for more information:
> Still... on a 40G total available disk space, tracker indexed 34G of
> Evolution emails... that's weird. What version of tracker are you
> running? Or are you indexing other filesystems?
>

Revision history for this message
C de-Avillez (hggdh2) said :
#6

What version of tracker are you running? What has it been configured to do? This is starting to sound like a bug...

Revision history for this message
Steve Reith (steve-reith) said :
#7

Hi - it is the standard app with the gutsy download : 063-0ubuntu3.
Did you see the web site: http://www.gnome.org/projects/tracker/ ? 063 -
was full of bugs - and I suspect 064 is no better. I have the "in the
box" config - as far as I can see it only searches evolution and my
testing shows that it only returns evolution hits from any query.

On Fri, 2007-12-28 at 17:00 +0000, hggdh wrote:
> Your question #20822 on Ubuntu changed:
> https://answers.edge.launchpad.net/ubuntu/+question/20822
>
> Status: Open => Needs information
>
> hggdh requested for more information:
> What version of tracker are you running? What has it been configured to
> do? This is starting to sound like a bug...
>

Revision history for this message
Steve Reith (steve-reith) said :
#8

I am still having severe problems daily with trackerd. Currently System Monitor shows trackerd using close to 1.2 GB out of 1.2 GB of ram. Swap is nearly at 100 %. System is nearly unusable - hard disk has light on continuously - vmstat shows system with with constant si so - very high wait states - apparent need to disable trackerd until application is fixed

regards,

Steve Reith

Revision history for this message
Best Andrea Corbellini (andrea.corbellini) said :
#9

System→Preferences→Session

Revision history for this message
Steve Reith (steve-reith) said :
#10

Thanks,

Steve Reith