Help - Search - Members - Calendar
Full Version: Upgrading Ensim 4.0.2 or 4.0.3 to 4.0.4
The Planet Forums > Control Panels > Ensim > Ensim HOWTOs
Pages: 1, 2, 3, 4, 5, 6
EV1S-Scott
1. Disclaimer
Ensim has stated that versions 4.0.2 through 4.1.0 are End of Maintenance as of 07/2007 and End of Life as of 12/2007
Please note - ThePlanet can no longer provide upgrade support for Ensim Pro 4.x versions.

EV1Servers is providing this information in order to help their customers upgrade their Ensim 4.0.2 or 4.0.3 servers to Ensim 4.0.4. The Ensim installer will also upgrade the server to Red Hat Enterprise 4. During this process, a majority of the operating system will be upgraded and replaced. ThePlanet-EV1Servers cannot assume responsibility for any lost data or unexpected downtime that may result from using this guide.
---------------------------------------------------------------------------------------
2. Introduction

Ensim has recently released version 4.0.4 of their Webppliance software for the RedHat Linux environment. Part of the upgrade to Ensim 4.0.4 includes upgrading the underlying operating system to RedHat Enterprise Linux 4. As a result, the upgrade itself is very time consuming, taking at least 1.5 hours on an idle box. A production box should take several hours to fully upgrade. Also, during the upgrade, the kernel will be upgraded to a 2.6 RHEL4 kernel, so expect the server to reboot itself at one point.

Please keep in mind that you must be running Ensim 4.0.2 or Ensim 4.0.3 to upgrade to Ensim 4.0.4. If you are not running either of these, please upgrade to them first.

This guide is intended to be used along side the guide provided by Ensim. I suggest that you download and thoroughly read both guides. Ensim's guide can be found here .

Of course, you'll want your Ensim control panel working relatively well. Make sure you can at least log in to Ensim, list the sites and basic functionality works. If it doesn't, I suggest you open a ticket with EV1Servers and get us to resolve that first.

Due to the nature of this upgrade, we strongly urge our customers to make backups in the event that something goes terribly wrong with the upgrade program and a restore is needed.
---------------------------------------------------------------------------------------
3. Make Backups

Can't say this enough folks. The Ensim upgrader will be upgrading Ensim and the underlying operating system. While we have not observed any catastrophic failures in the upgrade script provided by Ensim, we cannot guarantee it won't happen. We strongly urge you to make backups. One possible method is to use the backup utility provided by Ensim in their control panel. If you're unable to make backups using this tool, please open a ticket with EV1Servers describing the problem and asking us to investigate it
---------------------------------------------------------------------------------------
4. Make sure you can access the repository

We currently have the repository set up to only provide access for servers on the EV1Servers network. However, considering how often we're bringing new subnets online with servers, there's a possibility we left your subnet out. Therefore, you should ping the repository to make sure you have access. You can do it like this:

CODE
[~] $ ping rhel4.ev1servers.net

PING rhel4.ev1servers.net (207.218.247.251) 56(84) bytes of data.

64 bytes from rhel4.ev1servers.net (207.218.247.251): icmp_seq=1 ttl=62 time=0.259 ms

64 bytes from rhel4.ev1servers.net (207.218.247.251): icmp_seq=2 ttl=62 time=0.351 ms

64 bytes from rhel4.ev1servers.net (207.218.247.251): icmp_seq=3 ttl=62 time=0.243 ms

64 bytes from rhel4.ev1servers.net (207.218.247.251): icmp_seq=4 ttl=62 time=0.253 ms



--- rhel4.ev1servers.net ping statistics ---

4 packets transmitted, 4 received, 0% packet loss, time 3002ms

rtt min/avg/max/mdev = 0.243/0.276/0.351/0.046 ms


If you don't see a consistent set of ping responses, like the ones above, then you should submit a ticket asking us to add your subnet
---------------------------------------------------------------------------------------
5. Remove Packages with Unresolvable Dependencies

There are some packages which ship with RHEL3 that currently have unresolvable dependencies in RHEL4. As a result, you'll need to remove these packages before beginning the upgrade process. Servers delivered by EV1S include two such packages: 1) gtkglarea; and, 2) ethereal. In order to remove these, please execute this command:

CODE
[root@box root]# rpm -e gtkglarea ethereal


If rpm complains that a package isn't installed, just ignore it. It will have removed the other package anyways. You should also remove any other packages that you've added since getting the server that you feel may have dependencies to RHEL3 specific package versions

Also, this is a good time to go through the "Before you upgrade" section of the Ensim manual.
---------------------------------------------------------------------------------------
6. Begin installation

At this point, it's time to start the installation. First, we suggest that you run the installer inside of a screen session. An introduction to screen can be found here. To start screen, just type use command:

CODE
[root@box root]# screen


Your session might blink as screen loads and then your set. Should you ever get disconnected and need to pick up where you left off, just run this command:

CODE
[root@box root]# screen -rx


Now, download the Ensim installer script:

CODE
[root@box root]# wgetftp://ftp.ensim.com/apt/ensim/LWP/4.0.4/12.rhel.4ES/tools/ensim-installer.sh


Start the installer by using this command. Make sure to put a valid E-Mail address into the command.

CODE
[root@box root]# sh ensim-installer.sh --cdrom=n --email=address@domain.com -u [url=http://rhel4.ev1servers.net/updates]http://rhel4.ev1servers.net/updates[/url] -ohttp://rhel4.ev1servers.net/base


Let the installer do its work. It'll leave additional information in /root/.webppliance/upgrade.log. When Ensim is ready, it'll reboot the server so that it can begin to use the new 2.6 kernel. After the reboot, it'll do maintenance on each of the sites to make sure everything is squared away for Ensim 4.0.4. This process takes some time, and you can watch its progress in /var/log/messages. You can tell it's done when this line appears in /var/log/appliance/setup.log:

CODE
[success] POST_MAINTENANCE: Successfully left post maintenance mode


At this point, it'd be a good idea to carry out the "Post-upgrade tasks" found in the Ensim upgrade guide.

Please note - ThePlanet can no longer provide upgrade support for Ensim Pro 4.x versions.
If you have any issues with the installer, please visit the Ensim section of the EV1Servers Forums. The forums are usually full of people who are prepared to provide some advice. Finally, there's also the forums provided by Ensim at forums.ensim.com.
---------------------------------------------------------------------------------------

If anyone sees any typos or sees that this guide may need some modification, email us at systems-unix [at] theplanet.com

Thank you and good luck!

Scott O'Neil
Level 2 Technician, EV1Servers

---------------------------------------------------------------------------------------
12/2006 - Edited by Kevin-Ev1 Sr. Unix Administrator
  • Scott O'Neil is no longer a employee of ThePlanet-Ev1servers. He's moved onto other adventures. We wish him the best! He was employed here at ThePlanet-Ev1servers when he wrote this so we are just changing the point of contact information. If the guide needs modification, please email systems-unix [at] theplanet.com
  • rhel4.ev1servers.net is moving to 207.218.247.251 from 67.15.60.108, sorry for any downtime this may cause you. Links and URL will be the same.
---------------------------------------------------------------------------------------
06/2007 - Edited by Aaron - Product Manager
  • Please note - ThePlanet can no longer provide upgrade support for Ensim Pro 4.x versions.
jbeck
I have attempted to upgrade my box (4.0.3 => 4.0.4).....

While the server was upgrading it went into a file system check. Now everythign is wrecked. My services will not run and I get the following error message when I try to run the upgrade scrypt again.

CODE
sh ensim-installer.sh --cdrom=n --email=myemailaddress -u [url="http://rhel4.ev1servers.net/updates"]http://rhel4.ev1servers.net/updates[/url] -o [url="http://rhel4.ev1servers.net/base"]http://rhel4.ev1servers.net/base[/url]

08/26/05 04:09:10 : Ensim Installer

08/26/05 04:09:10 : Logging to file /tmp/install.log.2504. Final logs will be appended to /var/log/ensim/installer.log

08/26/05 04:09:10 : Invoked with arguments: --cdrom=n --email=myemailaddress -u [url="http://rhel4.ev1servers.net/updates"]http://rhel4.ev1servers.net/updates[/url] -o [url="http://rhel4.ev1servers.net/base"]http://rhel4.ev1servers.net/base[/url]

ensim-installer.sh: line 1306: /usr/bin/ensimchecker: No such file or directory

error: package ensimchecker is not installed

08/26/05 04:11:32 : Upgrading/Installing Ensim Control Panel: FAILED


my /var/log/ensim/installer.log contains

CODE
08/26/05 00:08:16 : Ensim Installer

08/26/05 00:08:16 : Logging to file /tmp/install.log.30481. Final logs will be appended to /var/log/ensim/installer.log

08/26/05 00:08:16 : Invoked with arguments: --cdrom=n --email=myemailaddress -u [url="http://rhel4.ev1servers.net/updates"]http://rhel4.ev1servers.net/updates[/url] -o [url="http://rhel4.ev1servers.net/base"]http://rhel4.ev1servers.net/base[/url]

08/26/05 00:09:05 : Running system check...

WEBppliance Pre-install/upgrade Checker 1.0



Environment:

   WEBppliance 4.0.3-22.rhel.3ES

   Red Hat Enterprise Linux ES release 3 (Taroon)



kernel       check kernel version                                     [   OK  ]

selinux      check for selinux enabled                                [   OK  ]

consistent   check if all sites are in a consistent state             [   OK  ]

rpm          check the version of rpm                                 [   OK  ]

dtml         check webppliance .dtml files                            [   OK  ]

logrotate    check /etc/logrotate.conf                                [   OK  ]

mysql        check for correct mysql rpms and binaries                [   OK  ]

indexshtml   check /var/www/html/index.shtml                          [   OK  ]

hostname     check the hostname and IP address                        [   OK  ]

perl         check for working perl                                   [   OK  ]

pwdb_locks   check for lock files.                                    [   OK  ]

extras       check /usr/lib/opcenter/fastcgi/extras/                  [   OK  ]

appsendmail  check appliance sendmail configuration                   [WARNING]

            the source for the appliance sendmail configuration is modified

            the appliance sendmail configuration is modified

run_as_root  check that the installer is running as root.             [   OK  ]

sitesendmail check site sendmail configuration                        [   OK  ]

zonetmpl     check /usr/lib/opcenter/bind/named_conf_zone.tmpl        [   OK  ]

mysqlon      ensure the mysql server is running, if installed         [   OK  ]

suexec       check /usr/sbin/suexec                                   [   OK  ]

net          check for non-loopback net device                        [   OK  ]

lang         check default language                                   [   OK  ]

iris         ensure the iris skin add-on is not installed             [   OK  ]

default      check default plan                                       [WARNING]

            the default plan is modified

postgreson   ensure the postgres server is running                    [   OK  ]

vacuum       run vacuum analyze on the postgres database              [   OK  ]

daemons      check if the database servers can be restarted           [   OK  ]

mount        check for bind mounts                                    [   OK  ]

usrjava      ensure there is no broken /usr/java symlink              [   OK  ]

htaccess     check for non-frontpage .htaccess files                  [   OK  ]

sqmail       check sqmail configuration files                         [   OK  ]

deleted_sites check for partially deleted or empty sites.              [   OK  ]

domainmap    check site to domain map                                 [   OK  ]





DOMAINS WITH ISSUES:



   none



APPLIANCE ISSUES:



   modified /etc/mail/sendmail.cf

   modified /etc/virtualhosting/plans/default/apache

   modified /etc/virtualhosting/plans/default/mysql

   modified /etc/virtualhosting/plans/default/ssh

   modified /etc/virtualhosting/plans/default/telnet

   modified /usr/lib/opcenter/sendmail/install/sendmail.mc



RECOMMENDED ACTIONS BEFORE UPGRADE:



   backup any modified files



RECOMMENDED ACTIONS AFTER UPGRADE:



   restore or merge changes in files you backed up



SUMMARY for srv01.myserver: 0 error(s), 2 warning(s), 0 domain(s) with issues



08/26/05 01:46:25 : Upgrading/Installing Ensim Control Panel: SUCCESSFUL



If I get a copy of /usr/bin/ensimchecker and try again do you experts think it will get fixed?

I am running out of time to get this running.. the sun is comming up now :-(
EV1S-Scott
Hi jbeck,
I suggest you open a ticket and PM me or E-Mail me your ES/RS number or IP address. I'll provide help where I can.

Thank you,

Scott O'Nel
Level 2 Technician, EV1Servers
scott [at] ev1servers.net
Catalyst
BE FOREWARNED!

Ensim 4.0.4 DOES NOT include php-pear! Any sites you have requiring it (which you'll probably have) will fail after the upgrade.

RHEL4 php moved PEAR out to php-pear ... you'll need to use ensimpkg or something to propogate it to your virtual sites.
Catalyst
jbeck, check out /root/.webppliance/upgrade.log ... That should tell you what's going on.

Mine didn't correctly obsolute rh-postgresql ... and there was a libcurl issue ... and the usual MySQL funkinees (had to install mysql-shared-compat & mysql-shared postupgrade).
jbeck
scott has been investigating/fixing my problem and thinks my problem was caused by my filesystem failing durring the upgrade...

I don't know what all he has done so far but my server is mostly working now.

excellent service I must say!
Seeker
is there a way to force a file system check? maybe running that before one attempts to upgrade is a solution? if machine is due to go bad it will go bad before upgrade and not during?
wshawn
You can run a system check by running the installer and then saying No when it askes to continue at the end of the check...


For everyone's information: the RHEL4 repository has been taken down and as of the response from the trouble ticket and online chat just finished there is no expected ETA for when it will be back up..

Make sure to ping the server per scotts how to and if it is not there your upgrade will of course bomb...

Mine just did, but fortunately it had not done much outside of

QUOTE
tail -f /root/.webppliance/upgrade.log
ls: /boot/*.img.orig: No such file or directory
boot_img_to_img rollback done.
data.old rollback in progress.
Starting postgresql service:                               [  OK  ]
/var/lib/pgsql/data.old can be safely deleted.
data.old rollback successful.
Stopping crond:                                            [FAILED]
Starting crond:                                            [  OK  ]
Starting bandwidth_manager:                                [  OK  ]
bandwidth manager restarted.


QUOTE
less /root/.webppliance/upgrade.log

Boot image rollback, not required.
data.old rollback, not required.
Bandwidth manager start, not required.
Boot image rollback, not required.
data.old rollback, not required.
Bandwidth manager start, not required.
Boot image rollback, not required.
data.old rollback, not required.
Bandwidth manager start, not required.
Boot image rollback, not required.
data.old rollback, not required.
Bandwidth manager start, not required.
               Performing Pre Upgrade tasks.
               -       Rule19 mode activated.
Spam Bayes DB Pre Upgrade Tasks

-----  Deleted a bunch of Spam db stuff

Updating Spam Database.
Backing up appliance database
Database has been successfully backed at /var/lib/pgsql/backups/db_dump.gz
Stopping postgresql
Stopping postgresql service: ESC[60G[  ESC[0;32mOKESC[0;39m  ]
Moved existing database to /var/lib/pgsql/data.old
               -       Moved appldb successfully.
               -       /boot/initrd-2.4.21-15.EL.img
/boot/initrd-2.4.21-15.ELsmp.img
/boot/initrd-2.4.21-4.0.1.EL.img
/boot/initrd-2.4.21-4.0.1.ELsmp.img have/has been renamed.
               -       Resolving rpm dependencies and conflicts.
               kernel-2.4.21-15.EL
kernel-2.4.21-4.0.1.EL
kernel-smp-2.4.21-15.EL
kernel-smp-2.4.21-4.0.1.EL kernel(s) will be erased.
ESC[60G[  ESC[0;32mOKESC[0;39m  ]^MStopping crond: ESC[60G[  ESC[0;32mOKESC[0;39m  ]
               Installing yum.
               -       Creating /etc/yum.conf.ensim.
               -       /etc/yum.conf.ensim successfully created.
               Upgrading the box, this may take time.
http://rhel4.ev1servers.net/base/repodata/repomd.xml: [Errno 4] IOError:
Trying other mirror.
Cannot open/read repomd.xml file for repository: os
failure: repodata/repomd.xml from os: [Errno 256] No more mirrors to try.
Setting up Upgrade Process
Setting up Repos
Rechecking for yum's pid
Yum exited successfully.
Cleaning up Packages
0 packages removed
ls: /boot/*.img.orig: No such file or directory
boot_img_to_img rollback done.
data.old rollback in progress.
Starting postgresql service: ESC[60G[  ESC[0;32mOKESC[0;39m  ]
/var/lib/pgsql/data.old can be safely deleted.
data.old rollback successful.
Stopping crond: ESC[60G[ESC[0;31mFAILEDESC[0;39m]
Starting crond: ESC[60G[  ESC[0;32mOKESC[0;39m  ]
ESC[60G[ESC[0;31mFAILEDESC[0;39m]^MStarting bandwidth_manager: ESC[60G[  ESC[0;32mOKESC[0;39m  ]
bandwidth manager restarted.


Seems to have rolled back to the privious stage with nothing bothered becuase the mirror was down...

All sites still functioning as in the Panel as they were not touched yet.
EV1S-Scott
Hi wshawn,
The RHES4 repository shouldn't be down. I've checked from my Ensim server and it appears to be working well. I'm afraid the technicians you spoke with just didn't know about the subnet issue that I had mentioned in the how to. Can you please PM me your RS or ES number and I will get this taken care of for you right away.

Thanks,

Scott
wshawn
I'm in the 67.15 subnet also... thought that would be included... PM on the way
wshawn
got it Scott.. thanks...
wshawn
7 hours later and we are finally finished with a server comming from 4.03 with 92 sites.

Plus side:

We ended up fixing all the screw ups in the file system from the 4.03 upgrade (causing a 30% increase in site sizes for no known reason)

Minus side:

Have found none as of yet. Even with errors spaced out throughout the upgrade everything seems to be running well. Just wish there was a lot more communication from the script. We ran multiple ssh connections with the update running in screen, while we watched 6 different log files update...

Thanks Scott and Ev1. Ensim should be paying you guys... icon_smile.gif
Daroz
QUOTE (wshawn)
Plus side:

We ended up fixing all the screw ups in the file system from the 4.03 upgrade (causing a 30% increase in site sizes for no known reason)


Q: Was this a fix by the upgrade, or was there something you guys had to manually go do to fix this?
JamesC
Well we have decided not to even try this upgrade ... we are going to move all our sites over to cPanel from the load difference we have seen and the ease of maintaining the server Ensim is just not an option for us anymore.

The last couple of upgrades caused just to much down time and even cost with server re-imaging. So its off to cPanel land from now on.
omnibus
Is there a simple way to import Ensim site on Cpanel?
Thank you.
wshawn
QUOTE (Daroz)
Q: Was this a fix by the upgrade, or was there something you guys had to manually go do to fix this?



The upgrade did it. As the file system was re written by the upgrade the softlinks were restored. On a typical website there are hundreds of megs of fluff in softlinks. If those become hard links then they get backed up with the rest of the site (instead of just the links).

Tho we are still probably running 6 gigs heavy that is much better than running 20 gigs heavy,
wshawn
QUOTE (JamesC)
Well we have decided not to even try this upgrade The last couple of upgrades caused just to much down time and even cost with server re-imaging. So its off to cPanel land from now on.


Have to agree with your point of view James, but as I have been told more than once (as I am probably the most vocal against Ensim) there is no perfect Control Panel and all have their problems.

I plan on needing anew server by years end, so at that time I will transfer over the critical (read that yellers) websites leaving them live on both serves for propagation. After all the sites are moved then I will reimage this drive to prestene (not to mention ahve a tech change the ribbon cables as they have both drives on the same chain making the slave SLOWWWW).

Enjoy Tennesse friend. I love it there when I can get there. Now back to our topic: Upgrade went pretty smooth for us. In fact the smoothest since Ensim 3.0
psyon
QUOTE
Error: Missing Dependency: kernel-source is needed by package tg3


I get the above error when attempting the upgrade. tg3 seems to be the ethernet drivers for the system.
theuruguayan
QUOTE (omnibus)
Is there a simple way to import Ensim site on Cpanel?
Thank you.


yes, the only problem with the import is the different database names, that can take you some work renaming the dbs..but that the only thing, then the rest is smotly..

glad you are doing good with your server james. any problems let me know..
Seeker
Questions:

Does EV1 offer reimage to Ensim 4.04?
If so, has anybody backed up their sites in Ensim 4.03 and succesfully load them into Ensim 4.04?

I only have 25 sites but i think I may just want to reimage and get it over with, either that or move to Plesk. (if that was the case can Plesk import Ensim sites/backup files? - dunno that may be more trouble than its worth.
Penguin
Ev1 currently don't offer a restore to RHEL4 and Ensim 4.0.4, as I can confirm from a very badly failed upgrade last night icon_sad.gif They will however restore you to RHEL4 so that you can manually install Webppliance, which I can confirm now works due to much appreciated help from Scott and some adjustments at EV1's end icon_wink.gif

If you back up sites, please read the comments in the Ensim forums - Ensim have apparently confirmed a bug that sites which don;t have majordomo enabled will not restore into 4.0.4, therefore ensure that this is enabled before you backup
Catalyst
That's weird, because four people offered to do it for me last week (re-image to RHEL4, with Ensim 4.0.4 installed). As a matter of fact, they wouldn't STOP offering even after I'd told them no...
psyon
Any suggestions on how to solve the tg3 dependancy problem?
Catalyst
QUOTE (psyon)
Any suggestions on how to solve the tg3 dependancy problem?
Stupid question, but have you installed kernel-source?
EV1S-Scott
Hi Psyon, All,
Eh, that tg3 problem is a little difficult to deal with. As far as I can tell, RHEL4 is using the same strategy that Fedora employed with respect to the kernel-source package. That is, Red Hat has deprecated the kernel-source package and expects the users to use the source RPMs. Well, that's fine and dandy, but it causes headaches for dependencies. Here's an excerpt from Red Hat's release notes:

QUOTE
In order to eliminate the redundancy inherent in providing a separate package for the kernel source code when that source code already exists in the kernel's .src.rpm file, Red Hat Enterprise Linux 4 no longer includes the kernel-source package. Users that require access to the kernel sources can find them in the kernel .src.rpm file. To create an exploded source tree from this file, perform the following steps (note that  refers to the version specification for your currently-running kernel):


That's from http://www.redhat.com/docs/manuals/enterpr...e-notes/es-x86/. The tg3 package is a special set of loadable modules, I forgot for what specifically. It automagically rebuilds itself against the kernel source in /usr/src, but if it's not there, the magic can't happen. Thus, the RPM dependency. The RPM dependency will need to be satisified to get the Ensim installer to cooperate.

psyon - Are you still having issues with the upgrade? If so, can you please submit a ticket, and place in the top of the ticket: "ATTENTION: Please let Scott O'Neil know about this ticket". Also, please pm me or E-Mail me your RS/EV1S number. I'm currently thinking of ways to get around this dependency issue. I'll prolly just build a kernel-source package myself. Or, there's a couple of other ways around it too. I'll be able to investigate this and provide more information tomorrow afternoon.

Thank you,

Scott O'Neil
Level 2
EV1Servers Technical Support
psyon
Thank you Scott.

I was aware of the removal of the kernel-source package already. I was hoping there was an updated tg3 RPM. I know Fedora Core 4 has a kernel-devel package. It just contains the headers and config files required to build modules and such against the kernel. I am not sure if RHEL 4 contains it though.

I don't want to submit a trouble ticket just yet. I would prefer to avoid having my server upgraded when I am not around to do testing and deal with errors afterwards.
HostDude
Upgrade to ensim 4.0.4 worked like charm. There was problem with gpan's php 4.3.11 showing dependency erros, I removed and reinstall php using up2date to correct. Upgrade went smoothly thereafter.

Thanks, Scott for nice job
psyon
Scott,
I was able to just remove the tg3 package and do the update. The 2.6.9 kernel used by RHEL 4 contains built in tg3 support.

I am battling some other issues now. Ensim is running its maintenance mode, and hoping that sorts it out. Will let everyone know.
psyon
Well, the upgrade is done, but not without hassle. I removed tg3 rpm, did the update, and the system rebooted. Kept my fingers crossed until I got ping replys back.

First thing wrong, my apr version was wrong. I do not recall updating it before, but I had 0.9.5 that was missing a function that apache and some stuff needed. Quickly deleted the 0.9.5 stuff, changed some symlinks, and apache was working, but no sites loaded, I just got the default Redhat page. Found out the httpd.conf had gotten overwritten, so I moved httpd.conf.rpmsave to httpd.conf, and that problem was solved. Tried to load my sites, and the php sites were broken. Turns out some libraries were needed in the virtual file systems. Webppliance was bailing out of the maintenance mode. I had to manually create some directories in the virtual file system. If anyones curious they were:
/home/virtual/siteX/fst/etc/netplug
/home/virtual/siteX/fst/usr/lib/tls/i486
/home/virtual/siteX/fst/usr/lib/tls/i586
/home/virtual/siteX/fst/usr/lib/tls/i686
/home/virtual/siteX/fst/lib/tls/i486
/home/virtual/siteX/fst/lib/tls/i586
/home/virtual/siteX/fst/lib/tls/i686

The /etc/init.d/postgresql script go overwritten to, and I had to change that back to a symlink to postgresql_app_init. After that mainenance mode ran fine. Everything worked except my Secure IMAP. Quick check of maillog said it could not load /usr/local/ssl/certs/imapd.pem. The file it turns out is in /usr/share/ssl/certs/imapd.pem. I havent found where to change the location yet, so I did hard links to the files, and that problem was solved.

The process was a pain in the @#!@#!@#!, but in the end, im upgraded.
Catalyst
I've mentioned the Certs problem in a few threads. It's actually looking for certs in "/usr/local/ssl/certs)" for some reason (everything's still in /usr/share/ssl/certs). Just "ln -s /usr/share/ssl /usr/local/ssl" and you're good.
fr4
Hi,

I have upgrade one server with no problem icon_smile.gif
In another server I got This Error icon_sad.gif :

CODE
Transaction Check Error:   file /usr/include/c++/3.2.3/i386-redhat-linux/bits/os

_defines.h from install of compat-gcc-32-c++-3.2.3-47.3 conflicts with file from

package libstdc++-devel-3.2.3-39

 file /usr/include/c++/3.2.3/limits from install of compat-gcc-32-c++-3.2.3-47.

3 conflicts with file from package libstdc++-devel-3.2.3-39

 file /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/libstdc++.a from install of comp

at-gcc-32-c++-3.2.3-47.3 conflicts with file from package libstdc++-devel-3.2.3-

39

 file /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/libstdc++.so from install of com

pat-gcc-32-c++-3.2.3-47.3 conflicts with file from package libstdc++-devel-3.2.3

-39

 file /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/libsupc++.a from install of comp

at-gcc-32-c++-3.2.3-47.3 conflicts with file from package libstdc++-devel-3.2.3-

39

 file /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/cpp0 from install of compat-gcc-

32-3.2.3-47.3 conflicts with file from package cpp-3.2.3-39

soleting: redhat-config-kickstart.noarch 0:2.3.19-1 with system-config-kickstart

.noarch 0:2.5.16-2


Have you any idea :confused: ?
Can I uninstall libstdc++-devel, cpp, gcc, virtualhosting-fst-develenv (dependencies that i found) ?

Thanks
Fr4
EV1S-Scott
Hi fr4,
That's an interesting conflict. The following specificies which RHEL release the packages are in:

libstdc++-devel-3.2.3-39 Red Hat Enterprise Linux ES (v. 3 for x86)
compat-gcc-32-c++-3.2.3-47.3 Red Hat Enterprise Linux ES (v. 4 for 32-bit x86)

It seems to me that compat-gcc-32 should be deprecating libstdc++-devel. But, I feel that this dependency is cropping up from a different dependency. You can't really remove those packages, removing virtualhosting packages is a recipe for trouble.

Would it be possible for you to submit a ticket so that I can read the entire output from yum on your server? Once you submit a ticket, please pm or email me your RS/ES# or the esceleation ID. I'll look into it and have some more information for you.

Thanks,

Scott O'Neil
Level 2 Technician
scott [at] ev1servers.net
Albo
Upgraded the OS without a problem, upgraded the new Ensim version, rebooted the machine, failed while running site-upgrade. An Ensim tech fixed that for me.

Updated the latest realese from rh (long list of rpms) running up2date -u.
Installed php-gd (Ensim forgot to include it I guess) because all vbulletin registrations were broken.

Now, I find out that pop3/imap authentication is not working. I can authenticate myself to po3/imap server using a system user (admin, root), but I cannot login using virtual accounts that ensim creates.

Anyone has any idea how to solve this?

Albo
EV1S-Scott
Hi,
While I'm not familiar with this particular issue, I'm confident that either I or Ensim can resolve it for you. I've updated your post in the RedHat OS section, can you please refer to it:

http://forums.ev1servers.net/showthread.ph...1928#post361928

Thank you,

Scott O'Neil
System Support Specialist
scott [at] ev1servers.net
EV1S-Scott
Hi All,
All taken care of, for some reason, the ensimvwh clause from the /etc/nsswitch.conf file was missing. I added it to the passwd line. Previously it read:

passwd: files

After the fix, it reads:

passwd: ensimvwh files

This missing clause crops up every now and then, it's so random and rare at this point that we've not invested the time to determine what is causing it.

Thank you,

Scott O'Neil
System Support Specialist
scott [at] ev1servers.net
Albo
One of those things that drive you nuts, and at the end the solution is as simple as changing one line in a configuration file.

Thanks Sott.
Albo
Albo
I will share my fix for GD, a graphics library used by php used on registration forms.

# 1. check to see if gd and php-gd is already installed
# rpm -q php-gd
php-gd-4.3.9-3.8

# rpm -q gd
gd-2.0.28-4


# 2. If any of them is not installed, install it.

up2date -u gd
up2date -u php-gd


# 3. Fix Ensim's apache script that is missing php-gd

vi /etc/virtualhosting/filelists/apache.sh

add this missing line for php-gd:

S,rpm:php-gd


# 4. Propogate the changes to all virtual filesystems

set_pre_maintenance
set_maintenance
set_post_maintenance
service webppliance restart

This should fix the problem.

Albo
Catalyst
Do the same for php-pear --- it's even more necessary.
ArtieMcD
Shouldn't the fix for the php-gd and php-pear be included in the apache.custom.sh file instead of apache.sh.... Since apache.sh clearly states:

CODE
#!/bin/bash

#

# [B]NOTE: DO NOT MODIFY THIS FILE (Please see below for more information)[/b]

#

# This file contains the initial information necessary to specify the

# contents of the filesystem template for a particular service; e.g.

# siteinfo, apache, ssh, etc. We understand that there will be users

# out there who would like to modify the contents of these filesystem

# templates; handling this in a way that is compatible with RPM's

# behavior during upgrade requires a second file that contains the

# modifications.

#

# First, understand that all this initial file does is spit out

# information about the proposed contents for the filesystem template,

# with each line containing one piece of information. Thus, [b]in order

# to modify this information, you can create a new file called

# <service>.custom.sh (e.g. siteinfo.custom.sh); the lines of

# information written by this first script will be piped through

# the <service>.custom.sh script, and the resulting output will be

# what is finally considered for the filesystem template. [/b]Here is

# a straightforward example of how to add and remove additional data

# (we will remove the 'which' rpm and add the 'xmlto' rpm):

#

# #!/bin/bash

#

# # grep out the which RPM from the input, passing the rest through

# /bin/cat | grep -v '^S,rpm:which'

# # add an additional line specifying the xmlto RPM

# /bin/echo "S,rpm:xmlto"



...


For example, we create a apache.custom.sh file like so and run through maintenance mode and we are all set.

CODE
#!/bin/bash

/bin/cat

/bin/echo "S,rpm:php-gd"
fr4
Hi scott,

Thanks for your post icon_biggrin.gif
But I have read it too late :'(
So I have removed all gcc package, and virtualhosting package (it have dependecies to cpp) and then i have relaunched upgrade and it seems to work perfectly icon_biggrin.gif

Thanks
Fr4
jorgece
Hi Scot,

Already updated a server using your directions. Thanks.
In another one, everything went ok till :

(from /root/.webppliance/upgrade.log)
Rechecking for yum's pid
Yum exited successfully.
Cleaning up Packages
0 packages removed
ls: /boot/*.img.orig: No such file or directory
boot_img_to_img rollback done.
data.old rollback in progress.
Starting postgresql service: [ OK ]
/var/lib/pgsql/data.old can be safely deleted.
data.old rollback successful.
Stopping crond: [FAILED]
Starting crond: [ OK ]
Starting bandwidth_manager: [ OK ]
bandwidth manager restarted.

And that's it.
Next is a mail message stating upgrade failed.

Any idea ?

TIA
EV1S-Scott
Hi jorgece,
There were probably some dependency errors. What does this command return:

CODE
grep -i error /root/.webppliance/upgrade.log


The dependency errors are scattered somewhere in the yum output. If you'd rather have me take a look at the problem, please feel free to submit a ticket and ask for it to be forwarded to Scott. Also, please pm me or E-Mail me the escalation ID if you choose this route.

Thank you,

Scott O'Neil
Systems Support Specialist
BoiTaiTui
I just upgraded ensim 4.0.3 to 4.0.4. The upgrade went smooth, since I have 0 site.

I update all the package using up2date and there is one package that cannot be update.

[root@ensim ~]# up2date -l

Fetching Obsoletes list for channel: rhel-i386-es-4...

Fetching rpm headers...
########################################

Name Version Rel
----------------------------------------------------------
postgresql-server 7.4.8 1.RHEL4.1 i386

The following Packages were marked to be skipped by your configuration:

Name Version Rel Reason
-------------------------------------------------------------------------------
kernel 2.6.9 11.ELPkg name/pattern
kernel-doc 2.6.9 11.ELPkg name/pattern
kernel-smp 2.6.9 11.ELPkg name/pattern
kernel-utils 2.4 13.1.66Pkg name/pattern

The following Packages are obsoleted by newer packages:

Name-Version-Release obsoleted by Name-Version-Release
-------------------------------------------------------------------------------
rh-postgresql-server-7.3.10-1 postgresql-server-7.4.8-1.RHEL4.1
[root@ensim ~]# up2date -u postgresql-server

Fetching Obsoletes list for channel: rhel-i386-es-4...

Fetching rpm headers...
########################################

Name Version Rel
----------------------------------------------------------
postgresql-server 7.4.8 1.RHEL4.1 i386


Testing package set / solving RPM inter-dependencies...
########################################
postgresql-server-7.4.8-1.R ########################## Done.
Preparing ########################################### [100%]

Installing...
1:postgresql-server ########################################### [100%]
Starting bandwidth_manager: [ OK ]
error reading information on service rhdb: No such file or directory
[root@ensim ~]#

any help would be greatly appreciated.
Catalyst
What you'll have to do is ...
CODE
rpm -qa | grep postgres
and make sure remove all the rh-postgres* RPMs, --force --nodeps, then reinstall them using up2date. Once you do that, then just
CODE
ln -s /etc/init.d/postgresql_app_init /etc/init.d/rhdb
so that some of the "service rhdb" operations will work.
RandomWire
Well, i tried to do the upgrade too. So here's my story..
on first run i got these errors:
---
Error: Missing Dependency: libpspell.so.4 is needed by package php
Error: Missing Dependency: libcurl.so.2 is needed by package php
Error: Missing Dependency: libcom_err.so.3 is needed by package php-imap
Error: Missing Dependency: libcom_err.so.3 is needed by package php
Error: proftpd conflicts with vsftpd
--

So i removed vsftpd and php rpms. I had gpan's rpms for php 5.0.3 working ok. I then installed php from up2date. re-run the update and went "well".
After the server restart things seemed to be working ok. I then updated php from up2date. I then checked the sites i have on the server and they give me this error

---
Fatal error: Call to undefined function: mysql_connect() in /var/www/html/includes/mysql_connect.php on line 2
---

While it seems like i have all the necessary rpms installed, i can't figure out what's wrong. Any ideas/suggestions?
EV1S-Scott
Hi RandomWire,
Seems like the php-mysql package isn't installed. Can you please check and let us know if it its installed.

Thanks,

Scott O'Neil
Systems Support Specialist
scott [at] ev1servers.net
RandomWire
[root@srv1 filelists]# rpm -qa | grep php-
php-imap-4.3.9-3.8
php-mysql-4.3.9-3.8
webppliance-mod_php-4.0.4-12.rhel.4ES
php-mbstring-4.3.9-3.8
php-pgsql-4.3.9-3.8
php-ldap-4.3.9-3.8
webppliance-mod_php-frontend-4.0.4-12.rhel.4ES
php-pear-4.3.9-3.8
php-4.3.9-3.8

It's there. i had already check that.
Catalyst
QUOTE
Error: Missing Dependency: libpspell.so.4 is needed by package php
Error: Missing Dependency: libcurl.so.2 is needed by package php
Error: Missing Dependency: libcom_err.so.3 is needed by package php-imap
Error: Missing Dependency: libcom_err.so.3 is needed by package php
Install krb5-libs, curl-devel & aspell-devel. Then you'll wanna make sure you have:
CODE
php-4.3.9-3.8

php-devel-4.3.9-3.8

php-gd-4.3.9-3.8

php-imap-4.3.9-3.8

php-ldap-4.3.9-3.8

php-mbstring-4.3.9-3.8

php-mysql-4.3.9-3.8

php-odbc-4.3.9-3.8

php-pear-4.3.9-3.8

php-pgsql-4.3.9-3.8
Then, once you run through maintenance, should be all good.
RandomWire
Done as you suggested. Still getting the mysql_connect() error though.
RandomWire
First of all i'd like to thank all of you that replied and gave me your suggestions.

The error was the extension_dir in php.ini. For those that installed Gpan's rpms remember to change this in

CODE
/home/virtual/FILESYSTEMTEMPLATE/apache/etc/php.ini


it should be /usr/lib/php4


I passed my box through maintenance and it didnt propagate to all sites. Edited their respective ini's manually to get them fixed. Maybe i forgot something. 7 hours straight and really tired! I ll get a look at it again some time tomorrow.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2009 Invision Power Services, Inc.