aidan's questions

I had one instance of the app (shame on me), running in the australia-southeast1 region yesterday.

Last night was a critical time for the app; people were relying on it working flawlessly, and then halfway through the event, the instance just disappeared. About 8 minutes later, a new instance was created and continued to serve traffic.

Is this something I should expect? Instances randomly disappearing and restarting? I realise I should have had a couple of versions running, but what if both versions restart at the same time?

The logs don't show any errors at the point the app (a standard, simple Rails app) restarts. The logs just have an 8 minute gap, followed by Puma booting up in the new instance.

CPU usage was constant, and rarely above about 5%. Memory usage was around 840MB, with a slight peak of 880MB just before the restart. "Disk Bytes" was a constant ~15KB/sec, with a large spike of 40MB/sec around the time of the restart.

app.yaml

entrypoint: bundle exec rackup -p $PORT
env: flex
runtime: ruby

handlers:
- url: .*
  script: main.app
  secure: always

automatic_scaling:
  min_num_instances: 1
  max_num_instances: 4
  cpu_utilization:
    target_utilization: 0.75

I'm running MySQL on a remote server, and it suddenly started rejecting all connections:

$ mysql -h 192.168.1.10 -u root -p
ERROR 1129 (00000): Host 'web' is blocked because of many connection errors; unblock with 'mysqladmin flush-hosts'

So, I try this flush-hosts command...

$ mysqladmin flush-hosts -h 192.168.1.10 -u root -p
mysqladmin: connect to server at '192.168.1.10' failed
error: 'Host 'web' is blocked because of many connection errors; unblock with 'mysqladmin flush-hosts''

I.e. it's blocking the very un-blocking tool it recommends. Am I doing it wrong, or will I have to resort to ssh/cpanel/physical access?

I was just browsing the system logs on my alestic Ubuntu 10.04 AWS EC2 production server, and found this:

daemon.log.1:Dec  9 22:45:42 server1 /etc/mysql/debian-start[865]: /usr/bin/mysqlcheck: Got error: 1045: Access denied for user 'debian-sys-maint'@'localhost' (using password: YES) when trying to connect
daemon.log.1:Dec  9 22:45:42 server1 /etc/mysql/debian-start[865]: FATAL ERROR: Upgrade failed

I'm assuming it came from a routine apt-get dist-upgrade. Is this something I need to address? It looks serious.

I have a simple LAMP stack on an EBS Ubuntu instance on AWS EC2.

It's been running well for a while now, but I'm now starting to run low on hard disk space.

What are my options?

I've already attached an EBS volume and moved logs etc to it, but the MySQL database continues to grow.

Do I move the database to the attached volume? I have serious concerns about this though. Database speed is currently the performance bottleneck in the LAMP stack, and moving the database to an EBS could incur a considerable performance hit (if the speed of moving files to and from the volumes is anything to go by?)

I need some help diagnosing a fault

I have a development server and a production server. They are (in the large part) identical in setup. Both are Ubuntu Server 10.04 LTS installs.

I turned on the MySQL slow query logs on both servers, and it works on the development server, but not the production server. On development, the slow query log fills up fairly quicly, but on production, the log file remains empty.

And idea how I can diagnose why it's not working on production? Are there any error logs that might contain the information?

There's nothing obviously relevant in /var/log/mysql/error.log, and /var/log/mysql/mysql-slow.log actually exists - it's just empty.

my.cnf:

log_slow_queries = /var/log/mysql/mysql-slow.log
long_query_time = 2

Thanks for any direction.

I'm taking steps to prevent someone (i.e. me) accidentally wiping the production server out.

I'm currently looking at SVN. I'm concerned that someone could accidentally issue the svn update command without specifying any files to update, and hence update the entire codebase to the head version.

Is there a way to prevent this from happening? Perhaps make svn an alias for a little pass-through script that checks inputs?

(I've already set up an alias for rm, to prevent people removing more than three files at a time, without seeing a confirmation prompt: alias rm='rm -I'. Are there any other little tweaks people make to their production servers, to reduce the risk of accidental catastrophe?)

I'm using the following to count the number of files in a directory, and its subdirectories:

find . -type f | wc -l

But I have half a million files in there, and the count takes a long time.

Is there a faster way to get a count of the number of files, that doesn't involve piping a huge amount of text to something that counts lines? It seems like an inefficient way to do things.

$ echo -e "test1\ntest2" > logfile

$ echo -e "test1\ntest2" >> logfile

$ cat logfile
test1
test2
test1
test2

$ rm logfile

$ crontab -e
* * * * * echo -e "test1\ntest2" >> logfile

$ sleep 160

$ cat logfile
-e test1
test2
-e test1
test2

Why am I getting the -e in the output? Both crontab and bash are using /bin/echo

I have exim running on a development server.

It's currently not able to deliver mail anywhere other than locally. This is fine though - I don't want to be able to accidentally spam our entire userbase. But it does make it difficult to see if the email system is working.

So, I want to capture all emails that exim tries to send (these emails are generated by a web-based CRM system on the same server), and store them in a shared mailbox (that I plan to make accessible to all developers via Dovecot/IMAP)

i.e. I want to redirect all outbound email to a local mbox.

Is this possible? I'm a complete exim beginner, and struggling with it.

Over time I've generated a proliferation of Ubuntu servers, and its starting to get tiresome keeping them all up to date.

If I SSH in a particular server, and notice it needs an update, I just fire off sudo screen -d -m apt-get -y dist-upgrade and forget.

This seems to be an inefficient and slightly reckless way of keeping my servers up to date.

Are there better solutions?

Also, this update process occasionally requires that the server be restarted. As a result, none of my servers have an uptime of more than a month. But other people are boasting that they've got uptimes in the region of years. How is this possible?

Cheers.

I used to mount a windows share in Ubuntu server, with an entry in fstab:

//data/SharedFolder /media/SharedFolder/ smbfs user,defaults,credentials=/root/.creds,uid=root,gid=root 0 0

/root/.creds is a text file with three lines, my username, password and domain.

Users on the ubuntu server could write to this mount, but then I upgraded to 10.04 and now only root can write. Regular users can still read though.

mount currently tells me:

//data/SharedFolder on /media/SharedFolder type cifs (rw,mand,noexec,nosuid,nodev)

How do I make it world writeable again?

Thanks

I had Ubuntu 9.10 x64 Desktop installed on a nettop I have (that I normally run headless), and yesterday I decided to take the plunge and update to 10.04.

So, I plugged in a screen and usb mouse/keyboard, booted up and set to work.

It was 1am, and it was telling me it had 3hrs left to install all the new packages, so I unplugged the screen and usb mouse/keyboard, left the box running, and went to bed.

This evening, I plugged it all back in again to check progress. It's asking if I want to remove obsolete packages. I do, but neither the mouse nor keyboard work!

I can access the box via SSH like I normally do; is there any way I can re-enable the keyboard from there?

I'm reluctant to restart the box (via ssh) mid-way through such a complicated upgrade.

Thanks for any help!

lsusb (with wireless mouse/keyboard receiver unplugged):

Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

lsusb (with wireless mouse/keyboard receiver attached):

Bus 004 Device 005: ID 045e:005f Microsoft Corp. Wireless MultiMedia Keyboard
Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

I'm trying to configure sudo to allow all users to restart apache without having to enter a password.

Security concerns aside, why isn't this working?

I added the line to /etc/sudoers: %admin ALL=NOPASSWD: /usr/sbin/apache2ctl

$sudo -l
User aidan may run the following commands on this host:
    (root) NOPASSWD: /usr/sbin/apache2ctl
    (ALL) ALL

$sudo /usr/sbin/apache2ctl
[sudo] password for aidan:

Thanks for any help.

I'm worried about dodgy php scripts or other malicious software on my linux server sending out spam. Or maybe I left an open relay

What are the ways to check I'm not sending any spam out?

I've tried a few standard Ubuntu instances on Amazon's Web Services (AWS), but I can never find the apache package using apt-get.

Thanks, Aidan.

Using Amazon's AWS, I created a 1 GiB volume. I attached it to a running instance and formatted it with ext3. I then mounted it and and used df -h to report its size. It's apparently 414G. I'm already using 2G of it.

What's going on here?

I started ntpdate running every night as a cronjob (on two servers). And every morning I have an email telling me the outcome. On one server it's correcting by about 0.8 seconds every day, which seems like quite a lot to me.

But on the other server, it's correcting by about -4 seconds every day. This sounds like a massive drift. Does this indicate anything wrong, like a low bios battery? (The server is quite old.)

Thanks.

I have multiple developers on one server, each with their own copy of the same perl codebase checked out of SVN.

I want each developer to have their own subdomain pointing to their own copy of the codebase.

I tried putting PerlSwitches inside , but that didn't seem to work. e.g.

<VirtualHost dev1.devserver.com>
    PerlSwitches -I/home/dev1/www -w -T
    <Location /webapp/>
        SetHandler  perl-script
        PerlHandler WebApp::Handler
    </Location>
</VirtualHost>

<VirtualHost dev2.devserver.com>
    PerlSwitches -I/home/dev2/www -w -T
    <Location /webapp/>
        SetHandler  perl-script
        PerlHandler WebApp::Handler
    </Location>
</VirtualHost>

Thanks.

We have a linux based development server which we connect to from our windows based desktops (via ssh, scp, samba (mapped network drive), etc)

We're doing web development in perl, using apache2, mod_perl and MySQL, and our code base is under SVN source control.

What is the best way to allow all (currently 5) developers to develop independently?

In the past I've experienced running multiple instances of apache (achievable by hacking apache's source code and compiling)

But perhaps it would be better to use virtualisation and effectively give each developer their own development server (seems a bit overkill for web development though)

And if I only run one instance of apache, what's the best way to redirect incoming request to the correct codebase? port? sub-domain? IP address of client?

Thanks

I set up an ubuntu server on our corporate network. Added an account for each developer and shared each home directory using samba. This all worked great until I had to add a new user - I can't make the new guy's home directory available via samba.

Normally, I "map network drive" in windows, using \\servername\username and the username and password. This works great for the existing users, but not the new user.

There's nothing in the smb.conf file that lists specific users or directories, and I've barely modified from its default settings. I've restarted the service and server. All the new user's files and directories have the same r/w permissions as everyone else's. The new user is a member of all the same groups as the existing members.

I tried installing the samba client to help diagnose the problem, but that seemed to remove the samba server.

Any ideas how I can diagnose the problem?

Thanks,

Aidan.