strigi uses up more than 25GB of space

[Scott Lewin]

I went to boot up my machine to find it can't boot up becasue my home directory was too full. I then went into console mode and deleted a ISO which gave me enough hard drive space to boot up Kubuntu. When I went to search for files larger than a GB to my surprise it was strigi eating up all my space.

After deleting strigi and all the files it created I gained over 25 GB of space and this is on a new install; only about a week. Strigi should not be taking up that much space and really should not be causing the desktop to fail to boot up.

Using Kubuntu 8.04 with KDE 3.5.9.

[Devlin (devlinsmailbox)]

I too can vouch for this problem. I have an 80GB drive, my personal files equal 20.7GB, I recieved message telling me my drive was 98% full. After scouring the system and deleting a VM and two iso's. I discovered the cause was the .strigi folderthat proved to be 50GB's! To be exact it was the contents of:

 ~/.strigi/clucene

[Scott Lewin (sclewin)]

Devlin wrote:
> I discovered
> the cause was the .strigi folderthat proved to be 50GB's! To be exact it
> was the contents of:
>
> ~/.strigi/clucene
This is the exact same place my 25GB+ was.

[Geoff (palmerg)]

Same here. On my 180GB /home folder Strigi has eaten 100MB! The more disk I free up, the more it consumes. In fact I can even watch the numbers change by looking at its status. The Index Size increases steadily.

After killing the strigi daemon, deleting ~/.strigi, and restarting it, the same thing happens. Half an hour ago Status showed;

Documents in queue 0
Documents indexed 125620
Index size 775 MB
Status idling
Unique words indexed 208012

It now shows;

Documents in queue 0
Documents indexed 216000
Index size 1370 MB
Status idling
Unique words indexed 208012

and that's with it set to index just three folders containing 1,600 files and occupying 530 MB of space!

Multiple indexing of the same files seems confirmed when I do a search. I get dozens of identical hits for the same long search string -- a search string that should point to just one document..

(System: Kubuntu 8.04, upgraded from 7.10. KDE release 3.5.9. Strigi version 0.5.7)

[Trebacz (david-trebacz)]

I am also having the exact same problem. Kubuntu 8.04 AMD 64 bit, upgraded from 7.10.

1177412054 - /home/david/.strigi/clucene/_1hy6oh.cfs
1177652972 - /home/david/.strigi/clucene/_1d63iq.cfs
9038646368 - /home/david/.strigi/clucene/_1imrjs.fdt
11612615881 - /home/david/.strigi/clucene/_6tva5.cfs
11617255594 - /home/david/.strigi/clucene/_dnqac.cfs
11620680808 - /home/david/.strigi/clucene/_khksv.cfs
11620866639 - /home/david/.strigi/clucene/_y5964.cfs
11636452659 - /home/david/.strigi/clucene/_rbezk.cfs
11640208846 - /home/david/.strigi/clucene/_14z35x.cfs
11642117756 - /home/david/.strigi/clucene/_1bsxjd.cfs

Stringi had multiple extremely large cfs files in the clucene directory. I delete them and they come back after a few days. There is actually very little on the hard drive of this machine -other than some backup created by BackupPC and several Samba shares. What in the world is Stringi indexing?

[markoloka (markoloka)]

Same here...
Home folder size 5.2gb.
Strigi folder size 3.6gb.
Now i have 90mb left before i get locked out.
Someone should be quick and do something about before more people can't login into to their kubuntu's.
I never even use strigi as it sucks. Kfind is 1000 times better.

[FriedChicken (domlyons)]

There are so many people who reported the same bug. Why did no one set this bug as "Confirmed"?

[Devlin (devlinsmailbox)]

FriedChicken wrote:
"There are so many people who reported the same bug. Why did no one set this bug as "Confirmed"?"

Upon attempting to change status Launchpad responded:
"You are not the bug assignee nor the maintainer of strigi (Ubuntu), and therefore cannot edit this bug's status. "

My response:
Um....It wouldn't let me.... :-)

[FriedChicken (domlyons)]

Hm, that's curious ...

[GrokIt (alan-ajinfosearch)]

I also had this problem twice. The first time it took up 100 GB in four files. I killed them and a few days later it made _five_ 25 GB files! I had to turn it off. Strigi not usable and now it is going to be the default for KDE 4?

[Oka Kurniawan (oka-kurniawan)]

I am having a similar problem.

Is there any way to solve this? so how should I free my disk space? can I just delete the .strigi folder?

[Scott Lewin (sclewin)]

Oka Kurniawan wrote:
> Is there any way to solve this? so how should I free my disk space? can
> I just delete the .strigi folder?
Most people seem to be not using Strigi, that is what I am doing. You
could try google desktop if you like.

--
Your friend,
Scott

Sent to you from a Linux computer using Kubuntu Version 8.04

[Benny (bm-cage)]

I confirm the problem. A clueless user had 1% of free disk space.
Running KDirStat showed the problem to be strigi.
The user had 5 Gb of data, the /home/user/.strigi directory has 100 Gb !

I would suggest to give this bug a high priority. Normal users don't know how to hunt for this with df or Kdirstat.

This is on an upgraded 7.10 install to 8.04, also with AMD 64 version of Kubuntu.

[Pablo (plberd-kubuntu)]

Hello, I never really used strigi long enough for it to make indexes that large. I discarded it because results aren't really helpful.
I've read your problems and... the only thing I can think of is that strigi is indexing it's own indexes. I don't have strigi installed anymore, but if someone does... could you try to configure it so it doesn't try to index anything in the ~/.strigi directory?

[Oka Kurniawan (oka-kurniawan)]

Hi Scott
Thanks, I am now using google desktop, and so far very satisfied. I uninstalled strigi and delete the .strigi folder.

[markoloka (markoloka)]

I got rid of it simply by killing strigi process and then deleting that uber fat .strigi folder. Now i use kfind.
Strigi is most useless program i have on my kubuntu 8.04.

[climatewarrior (gabrieljoel)]

I think this is a duplicate of https://bugs.launchpad.net/ubuntu/+source/strigi/+bug/256339 or vice-versa. Also this seems to be the upstream version of this bug http://sourceforge.net/tracker/index.php?func=detail&aid=1983277&group_id=171000&atid=856302.

[[e-mail] (tarahwheeler)]

I too have a huge strigi folder, and strigidaemon is hanging one of my processor cores. What do I do?

[Benny (bm-cage)]

Note that there is hope in the future: http://www.kdedevelopers.org/node/3573

"It regularly checks the available space on the home partition and suspends indexing if the space runs low (also very simple via KDiskFreeSpace. Using Qt/KDE is so damn great! You really can focus on the important stuff!)"

Let's hope the changes become the default in KDE

[Devlin (devlinsmailbox)]

Benny wrote:
> Note that there is hope in the future:
> http://www.kdedevelopers.org/node/3573
>
> "It regularly checks the available space on the home partition and
> suspends indexing if the space runs low (also very simple via
> KDiskFreeSpace. Using Qt/KDE is so damn great! You really can focus on
> the important stuff!)"
>
> Let's hope the changes become the default in KDE
>
>
I and I imagine many others would rather searching and indexing were
disabled by default. It has caused untold nightmares for newbies of
which I've had to fix. I am trying to promote the virtues of KDE but
when this feature (Strgi) is enabled by default reporting "out of disk
space" errors is not endering newbies to Linux at all. It even took me
off guard when I first came across this error.

Strigi off by default please, it will help lower calls to me of "what is
going on, my hard drive is so full I can't boot into KDE!"

[Harald Sitter (apachelogger)]

It is not on by default anymore. And will never be again, at least not on it's own. Indexing in KDE 4 is handled by nepomuk, abusing strigi as kind of backend.

[Jonathan Thomas (echidnaman)]

Strigi as used as a Nepomuk indexer should not have these sorts of issues any more in Kubuntu 10.04.

[bartman2589 (bartman2589)]

This bug has resurfaced in Kubuntu 10.10, I had Nepomum/Strigi set to only index a handful of folders which contained less than 10,000 files total, yet at the time of this post my strigi index was over 4.7GB and was reporting as having indexed over 470,000 files!!!