Some Websites Load Slow + Mysql Timeout

I'm hosting about 6 private websites on my server, and some have problems.

My biggest website is a vb forum with 30.000 members and 1.5 million posts.

When i browse with firefox i see that vb forum loads fast, but phpads which is integrated on the top and footer loads very slow. When i browse with IE, it takes like 10 seconds before the whole site is loaded and sometimes IE even crashes.

When i try to login on another website / admin panel (oscommerce), i get a mysql gone away error, that means timeout. But why does it show that? I ran mysqlreport and can't find anything why it should display mysql gone away.
# ./mysqlreport
MySQL 5.0.15-standard uptime 3 19:36:47 Mon Feb 5 11:08:10 2007

__ Key _________________________________________________________________
Buffer usage 78.94M of 96.00M %Used: 82.23
Write ratio 0.60
Read ratio 0.00

__ Questions ___________________________________________________________
Total 19.61M 59.47/s
Slow 9 0.00/s %Total: 0.00 %DMS: 0.00
DMS 7.17M 21.74/s 36.56

__ Table Locks _________________________________________________________
Waited 7.57k 0.02/s %Total: 0.05
Immediate 16.50M 50.03/s

__ Tables ______________________________________________________________
Open 1.27k of 1800 %Cache: 70.56
Opened 1.82k 0.01/s

__ Connections _________________________________________________________
Max used 47 of 500 %Max: 9.40
Total 2.62M 7.95/s

__ Created Temp ________________________________________________________
Disk table 6.77k 0.02/s
Table 635.25k 1.93/s
File 50.37k 0.15/s

I can't find any problems on the server which may lead to this strange problem.

The server load is normal, around 1 to 1.5. No weird errors in the apache error log or system log.

This strange behaviour started 2 weeks ago, and appears 2 to 3 time a week.

The system is running:
centos 4.3 64-bit
kernel 2.6.16.17 #1 SMP
directadmin 1.29.0
mysql 5.0.15
php 5.2
apache 2

Zend optimizer 3.2.0
Eaccelerator 1.95

One thing i noticed is that every night when the directadmin cron finished, the server acts normal again.

The directadmin cronjobs are:
* * * * * root /usr/local/directadmin/dataskq
2 0-23/6 * * * root echo 'action=vacation&value=all' >> /usr/local/directadmin/data/task.queue;
#5 5 * * 0 root /sbin/quotaoff -a; /sbin/quotacheck -augm; /sbin/quotaon -a;
20 4 1 * * root echo 'action=reset&value=all' >> /usr/local/directadmin/data/task.queue
0 4 * * * root echo 'action=check&value=license' >> /usr/local/directadmin/data/task.queue
10 0 * * * root /usr/local/directadmin/plugins/awstats/hooks/cgi-bin/awstats_updateall.pl now -awstatsprog=/usr/local/directadmin/plugins/awstats/hooks/cgi-bin/awstats.pl && echo 'action=tally&value=all' >> /usr/local/directadmin/data/task.queue