Educating the world

Our blog has over 10,000 readers a month

Plesk update causes makewhatis to fail

November 7th, 2012

During the Plesk 10.4 update the installer tries to “clean up” your platform. “Clean up” is Plesk code for “Break". The update deletes everything in /var/cache. While this area is reserved for caching, most of the applications that use it don’t expect their directory structure (that they created on installation) to disappear from under them. This was the case for the /etc/cron.daily/makewhatis.cron cron program which runs makewhatis which requires the directory /var/cache/man.

/etc/cron.daily/makewhatis.cron:

/usr/sbin/makewhatis: line 388: /var/cache/man/whatis: No such file or directory
chmod: cannot access `/var/cache/man/whatis’: No such file or directory
/usr/sbin/makewhatis: line 388: /var/cache/man/whatis: No such file or directory
chmod: cannot access `/var/cache/man/whatis’: No such file or directory
/usr/sbin/makewhatis: line 388: /var/cache/man/whatis: No such file or directory
chmod: cannot access `/var/cache/man/whatis’: No such file or directory
/usr/sbin/makewhatis: line 388: /var/cache/man/whatis: No such file or directory
chmod: cannot access `/var/cache/man/whatis’: No such file or directory

There are a couple of ways to fix this but the majority will possibly get overridden on the next update if Plesk doesn’t address this issue, which they probably won’t because they usually don’t!

The simplest thing to do is create the missing folder branch and check it is still there after the next update.

Log in to your Plesk server and issue the following command:

mkdir /var/cache/man

More Plesk fixes are available on Fixing problems caused by updating Plesk

Unable to generate a temporary class

November 2nd, 2012

If you use serialisation then IIS dynamically compiles the extra classes it needs to perform the serialisation and it uses the Windows temporary folders to store the class files definition and class compilation artefacts. If IIS can’t write to the temporary folder you will get this error message:

Server Error in ‘/application’ Application.
Unable to generate a temporary class (result=1).
error CS2001: Source file ‘C:\Windows\TEMP\1kcmvhly.0.cs’ could not be found
error CS2008: No inputs specified

Description: An unhandled exception occurred during the execution of the current web request. Please review the stack trace for more information about the error and where it originated in the code.

Exception Details: System.InvalidOperationException: Unable to generate a temporary class (result=1).
error CS2001: Source file ‘C:\Windows\TEMP\1kcmvhly.0.cs’ could not be found
error CS2008: No inputs specified

Source Error:
An unhandled exception was generated during the execution of the current web request. Information regarding the origin and location of the exception can be identified using the exception stack trace below.

Stack Trace:
[InvalidOperationException: Unable to generate a temporary class (result=1).
error CS2001: Source file ‘C:\Windows\TEMP\1kcmvhly.0.cs’ could not be found
error CS2008: No inputs specified
]

System.Xml.Serialization.Compiler.Compile(Assembly parent, String ns, XmlSerializerCompilerParameters xmlParameters, Evidence evidence) +1357
System.Xml.Serialization.TempAssembly.GenerateAssembly(XmlMapping[] xmlMappings, Type[] types, String defaultNamespace, Evidence evidence, XmlSerializerCompilerParameters parameters, Assembly assembly, Hashtable assemblies) +2809
System.Xml.Serialization.TempAssembly..ctor(XmlMapping[] xmlMappings, Type[] types, String defaultNamespace, String location, Evidence evidence) +366
System.Xml.Serialization.XmlSerializer..ctor(Type type, String defaultNamespace) +951
WebApplication1.LicenseManager.getLicenceRequestResult(String url) in D:\views\davidne_view\kernel\WIP\WebApplication1\WebApplication1\LicenseManager.cs:16
WebApplication1.LicenseManager.obtainLicenseKey(String version, String feature, String sessionid, String username) in D:\views\davidne_view\kernel\WIP\WebApplication1\WebApplication1\LicenseManager.cs:33
WebApplication1.LicensedSessionModule.OnPreRequestHandlerExecute(Object source, EventArgs e) in D:\views\davidne_view\kernel\WIP\WebApplication1\WebApplication1\LicensedSessionModule.cs:55
System.Web.SyncEventExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute() +80
System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously) +270

Version Information: Microsoft .NET Framework Version:4.0.30319; ASP.NET Version:4.0.30319.272

If IIS can’t create files in that folder then the application will fail at runtime so, you must give IIS write access to C:\windows\temp.

  1. Right click on C:\windows\temp select Properties.
  2. Select the Security tab.
  3. Click Edit
  4. When the Permissions for Temp dialgue box opens click Add.
  5. In the Enter the object names to select(example) box enter iis_iusrs.
  6. Click Check Names and the system will change the name to HOST\IIS_IUSRS. Click OK.
  7. Highlight IIS_IUSRS (HOST\IIS_IUSRS) and in the Permissions for IIS_IUSRS box, check all the security boxes in the Allow column. You can’t check Special permissions so ignore that one.
  8. Then OK all the way out.

It moans about changing permissions on system folders so just click Yes to that. Now you should be able to reload the page without a problem.

There is a way to get around it by recompiling the classes needed for serialisation but I’ll save that one for another time.

Setting Google Chrome's Google Maps extension's location

October 28th, 2012

I downloaded the Google Maps extension from the Web Store and installed it into Google Chrome. Opened a new tab and clicked the Google Maps button, it got me straight to a zoomed out picture of the United States of America. I couldn’t find any options inside Chrome’s settings or extension settings to change my default location.

Had a bit of a Google and found quite a few forum threads of people with a location mismatch. The main suggestion was to install the English (UK) version of Chrome (instead of the English (US) version which is the default version on the download site). The next most popular suggestion was to change the language settings to add English (UK) and increase it’s priority. I tried all of the suggestions and none of it worked.

Today, when I clicked the Google Maps extension button it took me to the USA as usual but I noticed on the right information panel it said “United States Not your current location? Correct it". So I clicked the “Correct it” link. I entered my location and Google Maps took me there. I closed the tab and reopened a new one, clicked the Google Maps extension button and Hey Presto I’m taken straight to my real location.

It’s so simple to change the default location and the actual setting is just in the right place. If you are out and about you can easily change to any location at any zoom level. Perfect.

Open relay caused by Plesk 10.4.4 update

October 17th, 2012

Another great shout out to the boys at Plesk. Installing patch 10.4.4 turned my box into an open relay that allowed spammers to use my host to send junk email.

I started to see the following appear in the process list. We were also finding that emails that were legitimately being sent were taking ages to come through. Logged in and checked the process list and I was very surprised to see:

... xinetd -stayalive -pidfile /var/run/xinetd.pid
... \_ /var/qmail/bin/qmail-smtpd /var/qmail/bin/smtp_auth /var/qmail/bin/cmd5checkpw
... \_ /var/qmail/bin/qmail-smtpd /var/qmail/bin/smtp_auth /var/qmail/bin/cmd5checkpw
... \_ /var/qmail/bin/qmail-smtpd /var/qmail/bin/smtp_auth /var/qmail/bin/cmd5checkpw
... \_ /var/qmail/bin/qmail-smtpd /var/qmail/bin/smtp_auth /var/qmail/bin/cmd5checkpw
... \_ /var/qmail/bin/qmail-smtpd /var/qmail/bin/smtp_auth /var/qmail/bin/cmd5checkpw
... qmail-send
...  \_ splogger qmail
...  \_ qmail-lspawn | /usr/bin/deliverquota ./Maildir
...  \_ qmail-rspawn
...  |   \_ /var/qmail/bin/qmail-remote.moved msa.hinet.net ljmbcc@yahoo.com.tw super.web1126@msa.hinet.net
...  |   \_ /var/qmail/bin/qmail-remote.moved pchome.com.tw ljmbcc@yahoo.com.tw 0929204890@pchome.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved ms3.hinet.net ljmbcc@yahoo.com.tw jacky.wu@ms3.hinet.net
...  |   \_ /var/qmail/bin/qmail-remote.moved live.co.uk co-operative.bank.plc.uk@co-op.account.aiert-wcc.cooperative4.host-8-coop.co
...  |   \_ /var/qmail/bin/qmail-remote.moved ms33.hinet.net ljmbcc@yahoo.com.tw aaba@ms33.hinet.net
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw s0925639306@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw t1014176@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw ce_0938468477@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved xuite.net ljmbcc@yahoo.com.tw dayaworks@xuite.net
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw moto231@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw a0932217273@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw a7824331@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw jolin1122kimo@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw mayeh888@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw sd1903@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw pink_ciw@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw kime20060719@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw eva7752@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw godhelll@yahoo.com.tw
...  |   \_ /var/qmail/bin/qmail-remote.moved yahoo.com.tw ljmbcc@yahoo.com.tw dunnini@yahoo.com.tw
...  \_ qmail-clean

Another symptom was that the secure log was registering loads of connects. You might get this on a busy server but ours isn’t.

tail -f /var/log/secure
Oct 15 13:04:14 s15427035 tcp-env[19544]: connect from 62.193.50.2 (62.193.50.2)
Oct 15 13:04:14 s15427035 tcp-env[19545]: connect from 177.0.160.130 (177.0.160.130)
Oct 15 13:04:14 s15427035 tcp-env[19546]: connect from 184.77.96.103 (184.77.96.103)

To cut a really long boring story short, Plesks 10.4.4 update added extra services to the /etc/xinetd.d:

cd /etc/xinetd.d
ls -l smtp*
-rw-r–r– 1 root root 288 Oct 15 13:03 smtp_ok // switch off
-rw-r–r– 1 root root 397 Oct 15 10:50 smtp_psa
-rw-r–r– 1 root root 290 Oct 15 13:03 smtps_ok
-rw-r–r– 1 root root 398 Oct 15 10:50 smtps_psa // switch off

Edit smtp_ok and smtps_ok and update the disable line to “yes". If the line doesn’t exist then create it.
Edit smtp_psa and smtps_psa and update the disable line to “no". If the line doesn’t exist then create it.

When you have made the changes, restart xinetd:

/etc/init.d/xinetd restart

The first thing to do is stop qmail so it can’t send any more junk mail.

/etc/init.d/qmail stop

Now that we’ve sealed the hole left by the Plesk 10.4.4 patch, we’ll see how much mail we need to clean up.

/var/qmail/bin/qmail-qstat
messages in queue: 7589
messages in queue but not yet preprocessed: 0

I looked for ages tying to figure out how to delete mail from the qmail queue and there doesn’t seem to be a way in the standard qmail tool set! So on to the internet and I found Qmail-Remove. This tool deletes everything in the qmail queue. It has a couple of different options to only delete certain emails but I’ll leave that as an exercise for the reader.

I knew that everything in the queue was junk so I could do a blanket delete:

qmail-remove -d

Check it’s all gone, then restart qmail:

/var/qmail/bin/qmail-qstat
messages in queue: 0
messages in queue but not yet preprocessed: 0

/etc/init.d/qmail start

SQL for beginners

October 4th, 2012

There are plenty of articles on how to get going with SQL (or MySQL) for beginners. However, there aren’t many articles which take you back to basics describing the building blocks used and help you understand the concept of what a database actually is.

This document will hopefully introduce you to the kinds of things that SQL is capable of. Each example has the command and the results so hopefully you can just read and understand it. I’m aiming this document at people who may come across SQL as part of someone else’s job and feel like they should know a bit about it, but don’t really know where to begin.

So let’s get going…

Structured Query Language or SQL is a computer language used to get stuff out of a database. To understand this statement we need to know what a computer language is and what a database is. Simple again! A computer language is just a set of commands that will make the computer do stuff and a database is multiple groups of stuff.

Now that we have got that out of the way, let’s begin. For this lesson we’ll be using MySQL, partly because it’s free and partly because there are some instructions here to set up MySQL and PHP on a test system.

To understand SQL you have to understand a little bit more about the structure of a database. The groups of stuff I mentioned earlier are called tables. A table is a grid of rows and columns. A database is a collection of tables. Now you have, everything you need to know! So let’s do an example we can all relate to.

We’ll create a table to hold information on people. Table names tend to be plural names like people, parts, events, etc so in this case people and not person because the table holds lots of persons! We collect some relevant information together: first name, sex, age. Then we’ll try to do some thing useful with them.

Before we create the people table we must have a think about the best way of storing the information in the table. Each attribute, field or column needs to have a type. First name is text, sex is an enumeration of male or female and age is an integer. You could make all columns text but we would like to do sums with the age column so we’ll need to use a column type that allows this.

In the early days of programming the commands were separated by new lines but as the languages got more complex people wanted to spread the code over several lines and so a semi-colon was used to denote the end of a statement.

CREATE TABLE people (
first_name VARCHAR(20)
,sex ENUM(’male’, ‘female’)
,age INT
);

The words in capitals are keywords, they are reserved for use by the computer language. i.e. you can’t have a table called table! Test columns need to have a maximum length so you need to specify it.

Now that we have a table let’s put some stuff in it:

INSERT INTO people(first_name, sex, age) VALUES
(’David’, ‘male’, ‘40′)
,(’Jon’, ‘male’, ‘30′)
,(’Tom’, ‘male’, ‘20′)
,(’Sophie’, ‘female’, ‘10′);

It’s only a small table so let’s list the contents.

SELECT first_name, sex, age FROM people;

Produces the output:

+------------+--------+------+
| first_name | sex    | age  |
+------------+--------+------+
| David      | male   |   40 |
| Jon        | male   |   30 |
| Tom        | male   |   20 |
| Sophie     | female |   10 |
+------------+--------+------+

Here you can definitely see the rows and columns structure I mentioned earlier. Each row is a complete set of information, that is there is an entry for every column. We didn’t put any restrictions on the table so we could have not supplied information for some of the columns if we wanted. In which case it might have looked something like this.

+------------+--------+------+
| first_name | sex    | age  |
+------------+--------+------+
| NULL       | male   |   40 |
| Jon        | NULL   |   30 |
| NULL       | NULL   | NULL |
| Sophie     | female |   10 |
+------------+--------+------+

Where NULL denotes the absence of data. In database terms NULL values are a bit of a bane. This special NULL symbol says the information is not provided and so we can’t assume anything about it. Most applications usually filter out the rows where there are gaps in the information.

Do you remember SETs from school? All of SQL is based on that idea. So lets look at our data and rephrase our question using SET-speak.

Show me all the information for a person where the person is over 21 is translated into

SELECT first_name, sex, age
FROM people
WHERE age > 21;

+------------+------+------+
| first_name | sex  | age  |
+------------+------+------+
| David      | male |   40 |
| Jon        | male |   30 |
+------------+------+------+

Show me all the first names of all the females is translated into

SELECT first_name FROM people WHERE sex = ‘female’;

+------------+
| first_name |
+------------+
| Sophie     |
+------------+

You may think that we are using the WHERE clause to filter the data to get results but what we are doing is far more subtle than that. We are creating new sets by dividing up existing sets. The sets may be added together in a union or subtracted from each other in an intersection. When we create our subset of data we can use it as the bases for another query. This can be shown in the following query:

SELECT first_name FROM people
WHERE age < 35
AND first_name IN (
SELECT first_name FROM people WHERE sex = ‘male’
);

The inner SELECT is the sub-query and produces a set of males names. That set of people is used with another set filter to reduce the set even further.

+------------+
| first_name |
+------------+
| Jon        |
| Tom        |
+------------+

The previous query could have been written as the following but quite often you’ll find that the data is not conveniently placed in the same table! it’s in another table or is the result of a calculation.

SELECT first_name FROM people
WHERE age < 35 AND sex = ‘male’;

Now that we understand a little more about the concepts of sets and creating them let’s look at doing something (more) useful.

We will start simple and count the number of males and females in the data set. To the average MySQL’er this will seem like an odd choice to talk about first but it will hopefully be obvious (later) that it is the only starting point.

SELECT COUNT(first_name), sex
FROM people
GROUP BY sex;

Produces the output:

+-------------------+--------+
| COUNT(first_name) | sex    |
+-------------------+--------+
|                 3 | male   |
|                 1 | female |
+-------------------+--------+

The GROUP BY command is used all the time to collapse sets of information so that they can be aggregated. In this case we are collapsing around the sex column. To the SQL engine that gathers the rows the GROUP BY effectively makes the data look like this.

+------------+--------+------+
| first_name | sex    | age  |
+------------+--------+------+
| ?????      | male   |   40 |
| ???        | male   |   30 |
| ???        | male   |   20 |
| ??????     | female |   10 |
+------------+--------+------+

There aren’t many aggregate operations we can do on text columns like first_name, so the information is meaningless and unpredictable. The data is treated as a separate set of each unique entry in sex. Conceptually the data set has been split into two:

+------------+--------+------+
| first_name | sex    | age  |
+------------+--------+------+
| ?????      | male   |   40 |
| ???        | male   |   30 |
| ???        | male   |   20 |
+------------+--------+------+

and

+------------+--------+------+
| first_name | sex    | age  |
+------------+--------+------+
| ??????     | female |   10 |
+------------+--------+------+

The GROUP BY will allow you to create aggregate operations. For example:

SELECT COUNT(first_name), sex
FROM people
GROUP BY sex;

shows the number of rows counted in each of the two groups of data, giving the result:

+----------+--------+
| COUNT(1) | sex    |
+----------+--------+
|        3 | male   |
|        1 | female |
+----------+--------+

or the total age and average for those sets.

SELECT SUM(age), AVG(age), sex
FROM people
GROUP BY sex;

giving

+----------+----------+--------+
| SUM(age) | AVG(age) | sex    |
+----------+----------+--------+
|       90 |  30.0000 | male   |
|       10 |  10.0000 | female |
+----------+----------+--------+

When the GROUP BY is applied to the whole table the GROUP BY keywords may be omitted but it is there implicitly. For example:

SELECT SUM(age), AVG(age) FROM people;

+----------+----------+
| SUM(age) | AVG(age) |
+----------+----------+
|      100 |  25.0000 |
+----------+----------+

This document introduces the concepts of sets, creating subsets and then treating a set as an entire entity and doing something useful. A table is one big set and next time we’ll look at joining tables together to see what matches and what’s missing.