Home / user-258663

Yves's questions

Martin Hope
Yves
Asked: 2019-06-16 00:38:36 +0800 CST
  • 2

I know that systemd is a mechanism of service management.

But I don't know why there are two config files: /etc/systemd/journald.conf and /etc/systemd/system.conf.

For example, in the system.conf, I see there is a line: LogLevel=info, and in the journald.conf, there are some lines:

#MaxLevelStore=debug
#MaxLevelSyslog=debug
#MaxLevelKMsg=notice
#MaxLevelConsole=info
#MaxLevelWall=emerg

They look very similar.

What's the difference between both of them?

systemd systemd-journald
Martin Hope
Yves
Asked: 2019-03-08 06:51:48 +0800 CST
  • 0

Today I try to use the way that I've known for several years to enter the Single User mode: enter the edit mode of grub and add single at the end of line of kernel and type F10.

I thought I could enter the Single User mode without typing password but to my surprise, it asked me to type the root password. I don't know why because I always used this way to rescue my linux system, for example, if I lost the root password, I could use this way to make a new root password.

But now things changed...

So Ubuntu has blocked this non-password single user mode?

password root single-user
Martin Hope
Yves
Asked: 2019-02-21 22:44:14 +0800 CST
  • 13

I'm learning how to use systemd notify. I'm thinking it is a kind of mechanism, which allows a process to send a notification to another process.

So, I tried to make a notification with the command systemd-notify --ready --status="hello". Then I got an error: No status data could be sent: $NOTIFY_SOCKET was not set. It seems that it needs a listener, just like the socket. But I don't know how to make a listener to receive this notification.

Also, I've known that the service of systemd has some different types, one of them is notify. The doc said, Type=notify: identical to Type=simple, but with the stipulation that the daemon will send a signal to systemd when it is ready.. So it seems that the service whose type is notify can send notification too, but I don't know how to use it either.

notification systemd
Martin Hope
Yves
Asked: 2018-12-27 08:28:37 +0800 CST
  • 0

I'm working on Ubuntu 16.04.3 and I need to create a file at the OS shutdown time.

After reading this link: https://unix.stackexchange.com/questions/39226/how-to-run-a-script-with-systemd-right-before-shutdown

I created a file named test.service at /lib/systemd/system/ and here is the content:

[Unit]
Description=test

[Service]
Type=oneshot
ExecStart=/bin/true
ExecStop=/home/test.sh

[Install]
WantedBy=multi-user.target

Here is /home/test.sh:

#!/bin/bash

touch /home/myfile

However, after rebooting the system, I cannot see any file named /home/myfile.

shutdown services systemd
Martin Hope
Yves
Asked: 2018-11-03 00:08:19 +0800 CST
  • 5

My server dumps because of huge system log files. I've heard that we could limit the size of system log by adding such a line size 100m into the file /etc/logrotate.d/rsyslog to do so.

Now my /etc/logrotate.d/rsyslog becomes as below:

/var/log/syslog
{
        rotate 7
        daily
        size 100m
        missingok
        notifempty
        delaycompress
        compress
        postrotate
                invoke-rc.d rsyslog rotate > /dev/null
        endscript
}

My question is: how does size 100m work with rotate 7?

Without size 100m, rotate 7 and daily together mean that the system log will be rotated daily and it will keep 7 days worth of backlogs. Now if I add size 100m and some daily log is bigger than 100m, will it rotate immediately? If so, does it mean that I may get 7 backlogs in a day if some daily log grows too much (more than 700mb)?

log rsyslog logrotate
Martin Hope
Yves
Asked: 2018-04-25 02:50:55 +0800 CST
  • 4

I have two Ubuntu system and they are exactly the same.

I execute several apt-get install XXX on one system and I can get the relative deb packages at /var/cache/apt/archives/. Then I copy these deb files to the other system at /home/me/archives/ and execute apt-get install ./*.deb.

I thought it should install the deb packages locally but to my surprise, it still redownload all deb files. I don't know why.

There are three things weird:

  1. When I execute apt-get install ./*.deb, lots of messages show up:

    Note, selecting 'python-rospkg' instead of './python-rospkg_1.1.4-100_all.deb'
Note, selecting 'python-serial' instead of './python-serial_3.0.1-1_all.deb'
Note, selecting 'python-service-identity' instead of './python-service-identity_16.0.0-2_all.deb'
Note, selecting 'python-setuptools' instead of './python-setuptools_20.7.0-1_all.deb'
Note, selecting 'python-sip-dev' instead of './python-sip-dev_4.17+dfsg-1build1_amd64.deb'
Note, selecting 'python-sip' instead of './python-sip_4.17+dfsg-1build1_amd64.deb'
Note, selecting 'python-six' instead of './python-six_1.10.0-3_all.deb'

  2. I try to install one of the deb packages, for example, apt-get install libwebp-dev_0.4.4-1_amd64.deb, but I always get the error:

    Reading package lists... Done
Building dependency tree
Reading state information... Done
E: Unable to locate package libwebp-dev_0.4.4-1_amd64.deb
E: Couldn't find any package by glob 'libwebp-dev_0.4.4-1_amd64.deb'
E: Couldn't find any package by regex 'libwebp-dev_0.4.4-1_amd64.deb'

  3. I've also tried to use dpkg -i *.deb to install these local deb packages, but it produced the error about missing some packages, then I executed apt install -f to get the missing deb package from the Internet, but the missing packages coming from the Internet and the original local packages are exactly the same...

In a word, my system cannot install local deb packages as expected.

package-management dpkg apt
Martin Hope
Yves
Asked: 2018-04-07 01:05:25 +0800 CST
  • 4

I'm trying to compile a C++ project and got an error: xxx.so no such file or directory. In this case, normally I execute the command apt-cache search xxx and I can usually find a package name like xxx-dev, so then I simply execute apt install xxx-dev and it will be OK.

But in fact I don't know exactly what apt install xxx-dev will do. When we execute apt install xxx-dev, we can get some information about what would be installed, etc, but I want to know something more.

If I execute apt install xxx-dev, will it install more things than I need? For example, the error at the beginning told me that xxx.so was missing, however, when I execute apt install xxx-dev, I can usually see that many packages will be installed. Are all of these necessary? Is there a way to check exactly what apt install xxx-dev will do or to check what a deb package contains?

software-installation package-management apt
Martin Hope
Yves
Asked: 2018-03-18 00:22:52 +0800 CST
  • 1

For some reason, I have to install some packages of Ubuntu 16.04 to Ubuntu 17.10. I've tried to apt update && apt install pkgname but there are always some dependency issues.

What I'm trying is that I get the source of some package with the command apt source pkgname.

I've found that I got three things:

pkgname.dsc
pkgname.orig.tar.gz
pkgname.debian.tar.xz

The .dsc file seems to contain some information of pkgname, the .orig.tar.gz file is the source code, but I don't know what the .debian.tar.xz file is.

For example, to install libpcl-dev, I did things as below:

apt source libpcl-dev && tar xzvf pcl.orig.tar.gz && cd pcl && mkdir build && cd build && cmake .. && make install && make.

It seems to work but I can't get what I need, which means that there isn't a file named libpcl-dev in /usr.

package-management debian apt
Martin Hope
Yves
Asked: 2018-03-13 17:42:41 +0800 CST
  • 2

I just found that there are two big files in my Ubuntu 16.04:

/run/log/journal/c840a48...(a very long name)...debfed5d/system.journal
/var/log/installer/cdebconf/templates.dat

My goal is to make the filesystem of Ubuntu as small as possible, so I'm thinking if I can remove these two files because they are all big enough (more than 10 MB).

It seems that they are two log files, so can I simply remove them? If they come from some processes, is it possible to disable the log system of these processes to prevent the two files to be generated?

filesystem log files
Martin Hope
Yves
Asked: 2018-02-13 01:05:06 +0800 CST
  • 33

As we all know, to configure the network, we can edit /etc/network/interfaces and execute /etc/init.d/networking restart.

Today I try to use Ubuntu 17.10 and I've found the network configuration had changed. Now we should edit /etc/netplan01-netcfg.yaml and execute netplan apply.

OK, this might be a bad question but I really want to know why Ubuntu did such a change. Is there a strong reason to explain all of changes like this one?

networking 17.10 netplan
Martin Hope
Yves
Asked: 2018-02-02 00:46:57 +0800 CST
  • 7

I'm working on Ubuntu 14.04 and I found something that I can't understand.

I execute the commands as below:

apt update && apt list --upgradable

I've found that bash is in the upgradable list.

For now the output of bash --version tells me that the version of bash is 4.3.11(1).

Then I do apt upgrade.

To my surprise, I do bash --version again and the output doesn't change at all. The version of bash is still 4.3.11(1).

So is this normal?, Some package can be upgradable while its version doesn't change?

package-management upgrade bash apt
Martin Hope
Yves
Asked: 2018-01-29 17:05:28 +0800 CST
  • 1

When I install the Ubuntu16.04, it seems that encrypting the home folder is possible.
enter image description here

I just want to know what it means. I've never selected it before. What will be different if I select it?

If we can encrypt the home folder, is it possible the encrypt any other directory? is it possible to encrypt a disk (for example the system is installed in /dev/sda and I want to encrypt the /dev/sdb)?

encryption partitioning
Martin Hope
Yves
Asked: 2018-01-24 23:36:17 +0800 CST
  • 4

Today I configured the PC of my company to setup a proxy and it worked because I can wget web page via proxy.

What I did is set http_proxy, https_proxy and ftp_proxy in the ~/.bashrc and in the ~/.wgetrc.

However, when I tried to curl www.google.fr, it blocked and timeout.

curl --proxy $http_proxy www.google.fr worked as expected.

Is it possible to make curl use the proxy automatically?

proxy
Martin Hope
Yves
Asked: 2018-01-23 21:51:35 +0800 CST
  • 1

I installed the Ubuntu-16.04.3-server-arm64 (hwe) on my ARM64 server.

enter image description here

Then I execute:

apt update && apt dist-upgrade && apt install ubuntu-desktop

Then I reboot. However, I can't see anything about desktop, I even can't see the ubuntu logo, which should have appeared before user login interface. But ssh on it is possible. So I think it's the problem of desktop.

I just checked the xorg log, there are some errors but I don't know what to do:

enter image description here

Here is the info about the graphics card:

enter image description here

16.04
Martin Hope
Yves
Asked: 2018-01-11 21:50:17 +0800 CST
  • 1

I'm trying to compile this c++ project: https://github.com/YvesZHI/FallingCode

I've successfully compiled it on my Mac OS, however, when I try to compile it on my ubuntu16.04, I get errors as below:

undefined reference to 'initscr'
...

which means that it can't find libncurses.

I've done: apt install libncurses-dev

I execute: pkg-config --cflags --libs ncurses and get the result:
-D_GNU_SOURCE -lncurses -ltinfo

g++ --version gives me

g++ (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.5) 5.4.0

I'm thinking if there is some error its makefile...

c++
Martin Hope
Yves
Asked: 2017-12-27 03:38:35 +0800 CST
  • 3

According to the man page of vmstat, 'swpd' means the amount of virtual memory used. As my understanding, for most cases it should be zero unless the physical CPU is out of space.

However, I just found that the value of 'swpd' of my server is not zero.
enter image description here

I don't know why. Even if I have more than 700 processes, the CPUs are almost 100% unused.
enter image description here

In total, the swap is used about 190M. It's not that big comparing with its total size which is 250G.

So does it mean that some processes will use the virtual swap even if the CPUs are free?

swap
Martin Hope
Yves
Asked: 2017-12-27 02:28:09 +0800 CST
  • 3

I'm using netstat to monitor my Ubuntu server and I've found things as below:

tcp     1     0     localhost:48604     localhost:11311     CLOSE WAIT 29511/st   
atic transf 
tcp     0     0     localhost:50810     X86U145:35723       ESTABLISHED 30019/nm 
ea2tfpose 
tcp     0     0     X86U145:42079       localhost:56772     ESTABLISHED 29535/ve    
ctor_map_lo 
unix 3     []        STREAM     CONNECTED    438037   29071/python
unix 3     []        STREAM     CONNECTED    478854   29071/python
unix 3     []        STREAM     CONNECTED    478852   29071/python
unix 3     []        STREAM     CONNECTED    487641   28978/gksudo
unix 3     []        STREAM     CONNECTED    436126   28978/gksudo
unix 3     []        STREAM     CONNECTED    473434   28978/gksudo

What do these lines mean?
unix 3 [] STREAM CONNECTED 438037 29037/python

networking