php 5.2.3 upgrade more issue

Warning: session_save_path() [
href=’function.session-save-path’>function.session-save-path
]:
open_basedir restriction in effect.

with error reporting set, blank page won’t give anything :-(

A little work need to be done since session not working properly after upgrade php from php 4.4.7 to php 5.2.3.

After doing install and uninstall php5 ports in freebsd 6.2 box. I can see that error.

Googling more to find some clue, I have one plus solution.

The idea is overwrite session_path

just make sure these lines exist in vhost directive :

php_admin_value open_basedir  /home/student

php_value session.save_path /home/student/session

make sure session in /home/student/session exist or make symlink to a directory.

It works now :-)

Upgrading to php5

Recent announcement from php.net about end support for php4 make me thinking about upgrading to php5.

All server already use latest php4 version php 4.4.7 with mysql support (4.0.27) and apache1.37

Upgrading start from student server :

1. Uninstall all related php4 ports

2. Install php5 ports

kongja# cd /usr/ports/lang/php5 && make clean
===>  Cleaning for apache-1.3.37_4
===>  Cleaning for php5-5.2.3
kongja# cd /usr/ports/lang/php5 && make install
===>  php5-5.2.3 : Your apache does not support DSO modules.
*** Error code 1

Stop in /data3/ports/lang/php5.
Continue reading

Finally Jeff commit SCHED_SMP work as SCHED_ULE

FreeBSD 7 release will get shaping up to be an awesome release :-)

From Jeff blog :

"Well 5 years ago this summer I stopped by evan‘s house in seattle’s university district and told him I was going to write a scheduler that saturday. Today I committed my SCHED_SMP work as SCHED_ULE and I feel it finally reflects my original hopes and intentions. Much of that time was spent getting the kernel locked well enough that we could exploit some affinity. However I have experimented with a great number of schemes and ideas in the process. For now I’m satisfied. I will take a little break and work on other projects."

Thanks for your efforts Jeff  😉

Need to order second CPU for preparing hehehehehe.

1610 temperature violation detected

New server arrive, HP proliant ML 350 G4 P . With four SCSI HD.

ready to replace our old email server :-)

Just a problem that arise with heatsink that prompt message " 1610 temperature violation detected "

sighhh…

wait more than 5 minutes for cold the machine has no impact.

A few result from google lead me to this post,

After remove "sticker like" stuff  between heatsink and cpu make everything goes normal.

simple thing but make me nuts for  few days :-(

Moral of story : keep fight :-)

 

Trying gstripe

New HP server came with 4 SCSI HD, need to use it for email server. Replace old Intel Pre Server host around 3000 email account.

I’ve been playing around with gconcat for ftp server, now wanna try gstripe :-)

jedimaster# gstripe label -v email /dev/da1s1d /dev/da2s1d /dev/da3s1d
Metadata value stored on /dev/da1s1d.
Metadata value stored on /dev/da2s1d.
Metadata value stored on /dev/da3s1d.
Done.

jedimaster# newfs -U /dev/stripe/email

jedimaster# mount /dev/stripe/email /email
jedimaster# df -H
Filesystem           Size    Used   Avail Capacity  Mounted on
/dev/da0s1a           10G     65M    9.5G     1%    /
devfs                1.0k    1.0k      0B   100%    /dev
/dev/da0s1d           52G    905M     47G     2%    /home
/dev/da0s1e           16G    3.9G     10G    27%    /usr
/dev/da0s1f           62G     20M     57G     0%    /var
/dev/stripe/email    427G    4.1k    392G     0%    /email

editing /etc/fstab

# Device                Mountpoint      FStype  Options         Dump    Pass#
/dev/da0s1b             none            swap    sw              0       0
/dev/da0s1a             /               ufs     rw              1       1
/dev/da0s1d             /home           ufs     rw              2       2
/dev/da0s1e             /usr            ufs     rw              2       2
/dev/da0s1f             /var            ufs     rw              2       2
/dev/acd0               /cdrom          cd9660  ro,noauto       0       0
/dev/stripe/email       /email          ufs     rw              2       2

jedimaster# echo ‘geom_stripe_load="YES"’ >> /boot/loader.conf

testing ::

jedimaster# time dd if=/dev/zero of=/email/test.img bs=1M count=1024
1024+0 records in
1024+0 records out
1073741824 bytes transferred in 21.451587 secs (50054191 bytes/sec)
0.000u 2.677s 0:21.46 12.4%     21+2974k 33+8192io 0pf+0w

hmmm….

is it scsi thing or I left something?

Links :

FreeBSD Gstripe

Interesting to see ZFS in action, I’ll try it too.