Questions[updatedb](ubuntu)

Whenever I run a command :

updatedb

I get the message :

File size limit exceeded (core dumped)

Why is that ?

I know that locate command utilizes a database created by a related utility, updatedb. Most Linux systems run this automatically once a day. However, you can update it at any time by just running updatedb from the command line as the root user.

update :

root@node1:~# ulimit -a

core file size          (blocks, -c) 0
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 0
file size               (blocks, -f) 2001
pending signals                 (-i) 23227
max locked memory       (kbytes, -l) 65536
max memory size         (kbytes, -m) unlimited
open files                      (-n) 1024
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) 8192
cpu time               (seconds, -t) unlimited
max user processes              (-u) 23227
virtual memory          (kbytes, -v) unlimited
file locks                      (-x) unlimited

For the longest time I have a huge performance issue in my Ubuntu VM, as every day at some time the cpu would be at a 100%, pointing to the fuse process that mounts the shared directories under /mnt/hgfs as the problem. To find out what was causing it I did

$ sudo lsof /mnt/hgfs/
[sudo] passord for myuser: 
COMMAND   PID   USER   FD   TYPE DEVICE SIZE/OFF    NODE NAME
find    20811 nobody    4r   DIR   0,45   327680 1644542 /mnt/hgfs/SharedData/Users/myuser/Pictures/Lightroom/Main Previews.lrdata/9

Followed by finding the parent of that find process:

$ sudo pstree -s -p 20811
systemd(1)───anacron(20328)───sh(20470)───run-parts(20472)───locate(20715)───updatedb.findut(20720)───updatedb.findut(20728)───su(20786)───sh(20809)───find(20811+

So it was a cron job that caused this by invoking updatedb. That's fine, I know updatedb has a config file at /etc/updatedb.conf and so I added /mnt/hgfs to the PRUNEPATHS variable and thought that would be it.

Turns out it was not. The same problem kept coming up.

OK, so grepping the /etc/cron.* directories I quickly discovered /etc/cron.daily/locate, and this time I took a look and discovered the script sets its own custom configuration, including PRUNEPATHS. It does not seem to consult /etc/updatedb.conf at all. This seems highly confusing to me. What is the point of a config file that is not used in the daily updates? Now I need to hack a system supplied file for this to work. That seems counter-intuitive and not something I am used to from Ubuntu.

Could someone explain the logic here, alternatively how this was supposed to work?

I am trying to build TensorFlow 1.6. in a Nvidia Jetson TX2, using a script located in: https://github.com/JasonAtNvidia/JetsonTFBuild/blob/master/BuildTensorflow.sh

In this script, "updatedb" is invoked (line 96) and I have a problem at this moment: BuildTensorflow.sh: line 96: /usr/bin/updatedb: cannot execute binary file: Exec format error

$ sudo updatedb /usr/bin/updatedb: 1: /usr/bin/updatedb: Syntax error: word unexpected (expecting ")")

Maybe these information can help;

$ sudo apt-get install mlocate
Reading package lists... Done
Building dependency tree       
Reading state information... Done
mlocate:i386 is already the newest version (0.26-1ubuntu2).
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

sudo apt-get install locate
Reading package lists... Done
Building dependency tree       
Reading state information... Done
locate:i386 is already the newest version (4.6.0+git+20160126-2).
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

$ dpkg -S /usr/bin/updatedb.findutils
locate:i386: /usr/bin/updatedb.findutils

$ dpkg --print-architecture
arm64

$ dpkg --print-foreign-architectures
i386

$ sudo gedit /etc/apt/sources.list
#--------------------------------------------------------------------- 
---------#
#                            OFFICIAL UBUNTU REPOS                             
#
#--------------------------------------------------------------------- 
---------#


###### Ubuntu Main Repos
deb [arch=amd64,i386] http://tw.archive.ubuntu.com/ubuntu/ xenial main 
restricted universe multiverse 
deb-src http://tw.archive.ubuntu.com/ubuntu/ xenial main restricted 
universe multiverse 

###### Ubuntu Update Repos
deb [arch=amd64,i386] http://tw.archive.ubuntu.com/ubuntu/ xenial- 
security main restricted universe multiverse 
deb [arch=amd64,i386] http://tw.archive.ubuntu.com/ubuntu/ xenial- 
updates main restricted universe multiverse 
deb [arch=amd64,i386] http://tw.archive.ubuntu.com/ubuntu/ xenial- 
proposed main restricted universe multiverse 
deb [arch=amd64,i386] http://tw.archive.ubuntu.com/ubuntu/ xenial- 
backports main restricted universe multiverse 
deb-src http://tw.archive.ubuntu.com/ubuntu/ xenial-security main 
restricted universe multiverse 
deb-src http://tw.archive.ubuntu.com/ubuntu/ xenial-updates main 
restricted universe multiverse 
deb-src http://tw.archive.ubuntu.com/ubuntu/ xenial-proposed main 
restricted universe multiverse 

$ head -n1 /etc/nv_tegra_release
# R28 (release), REVISION: 2.0, GCID: 10567845, BOARD: t186ref, EABI: 
aarch64, DATE: Fri Mar  2 04:57:01 UTC 2018

I am using Ubuntu 16.04

I really do not know how to solve this problem, any help/lead is appreciated.

Thank you in advance for your help!

Yesterday, while trying to unmount our (davfs2 mounted) work webdav server from my filesystem, I got device is busy.

Running sudo lsof | grep /mountpoint/of/davfs, revealed that the culprit was updatedb. Running man updatedb revealed that this is a command run to update a database for mlocate, most probably invoked by cron.

While I sometimes use locate, i don't need it to include all files on our work server (we got other search tools for that), and I definitely do not want this to interfere with me unmounting the filesystem.

I see that Ubuntu uses /etc/cron.{daily,weekly,monthly} for cron jobs. I also see that some things, like updatedb from mlocate, put their jobs there:

$ dpkg -L mlocate|grep y/m
/etc/cron.daily/mlocate

Somtetimes, I would like to disable some of these jobs (mlocate in this case). I can obviously sudo mv /etc/cron.daily/mlocate ~/cron.daily-dont-run and be over with it. However, two questions:

• Is this how this gets managed? Is there something like service to manage SysV init scripts or overrides such as described in Upstart intro, 10.41.1 Override Files? I would like to tell it "OK, I don't want you to start", but without touching what is installed (see next item for one of the reasons)
• If I delete and then reinstall mlocate, I'll again get the cron.daily entry. Also, I guess (is this true?) upgrades will probably not go smoothly (e.g. it will make a cron.daily entry or maybe possibly break the package if there's something in the cron entry that it depends on).

If you do, how do you manage / deal with the things such as the above?