Azure Managing Disk Space

Overview

In SFTP Gateway 2.001.00 and 2.001.01, we have identified an issues where the Linux disk can fill up over the course of a few weeks, if the server is under heavy load.

This article covers a workaround to keep disk usage under control.

Problem

The default size of a VM is 30 GB. Under light usage, this should be enough disk space.

The operating system and files for a default installation take up around 5 GB.

If the server is under heavy load, various log files can take up over 6 GB of storage, even when rotated. These logs include:

• /var/log/messages
• LDAP logs
• Cron logs

Also, there is a folder /ts-tmp that contains command-line output for task spooler commands. There are a few processes that use the task spooler.

This folder can grow indefinitely in size, because it's not a traditional log file managed by logrotate.

And under heavy load, the /ts-tmp can use up the remainder of your server storage within a matter of weeks.

Diagnosing and troubleshooting

To determine whether you are low on disk space, run the command:

sudo su
df -h

You should see the following output:

Filesystem                          Size  Used Avail Use% Mounted on
/dev/sda2                            30G   26G  4.3G  86% /

To see how much of this storage is being taken up by /ts-tmp, run the following commands:

cd /ts-tmp
du -sh .

Cleaning up files

You can immediately clear out old files, using this command:

find /ts-tmp -mtime +1 -type f -exec rm {} \;

This will search for any files in the /ts-tmp directory that are older than a day, and then delete them.

To automate this process, you can run a daily scheduled task. Run this command:

sudo su
crontab -e

And append the following entry:

0 1 * * * find /ts-tmp -mtime +1 -type f -exec rm {} \; > /dev/null 2>&1

Once a day, this will delete /ts-tmp files older than a day old.