Topic: upgrade to SME 8 from 7 http syntax error and invalid kernel operation

[rafa0786]

Hi

This has been posted before (Bug 7776 - Invalid KERNEL operation message after upgrade from v7.6 to v8)

http://forums.contribs.org/index.php/topic,50020.0.html

but i cant seem to find a solution for the problem, i have removed all custom templates before the upgrade yet i get the following error

httpd: Syntax error on line 88 of /etc/httpd/conf/httpd.conf: Cannot load /etc/httpd/modules/mod_imap.so into server: /etc/httpd/modules/mod_imap.so: cannot open shared object file: No such file or directory

udevd[565]: add_to_rules: Invalid KERNEL operation
udevd[565]: add_to_rules: invalid rule 'etc/udev/rules.d/51-by-id.rules:1

Is there a fix for this or should i do a fresh install and restore ?

[janet]

rafa0786

i cant seem to find a solution for the problem

You need to read those links more carefully & completely.

That forum link & bug report relate to required repos not being enabled,
AND
that during the upgrade from CD, the install process was not allowed to fully complete.

So you need to determine what the real cause of your issues are.
I suspect it is a case of not letting the upgrade fully complete.

Did you let the CD upgrade fully complete ?
Refer comment 18 of bug 7776 which says:
"The only real final message is the one that says something like, Remove the CD & reboot the server. It can take 15 minutes or longer to get from the 100% installed message to the remove CD & reboot message."

If you did not get that "Remove CD & reboot message", then you should run the upgrade from CD again, & ensure you let the process fully finish & only reboot when you see the above message. Do not be impatient, as it may take a long time.

Running the upgrade from CD & allowing it to correctly & fully complete should fix all your issues.

If you really did let the upgrade finish, then you can check the status of repos with
/sbin/e-smith/audittools/repositories
& show us the output.

Also read all this section of the FAQ for what is required re repos & how to fix repo problems
http://wiki.contribs.org/SME_Server:Documentation:FAQ:Section01#Yum_Updates

Note there is no point just checking & fixing the repos if you did not finish the upgrade correctly, you need to re-run the upgrade from CD.
Doing that will/should fix the repos as well as install all required el5 rpms.

[rafa0786]

Thanks for the reply

I read the links perfectly fine thank you..

The upgrade did complete successfully if it did not i would not have posted this.

I am also aware that there is an issue with the repo's but as i am new to Linux i am struggling with the solution.

I will run /sbin/e-smith/audittools/repositories tomorrow and follow http://wiki.contribs.org/SME_Server:Documentation:FAQ:Section01#Yum_Updates to try and fix it.

[janet]

rafa0786

The general direction of the forum report did mention interrupted upgrade, & ultimately led to a bug report & that bug report also identified that an incomplete upgrade process (or upgrade process interrupted before it really finished) was the underlying problem & cause of exactly the type of error messages you are receiving. Read bug 7776 for yourself. I deliberately tried the upgrade & rebooted when it said "100% complete" & that certainly caused many many errors of the type you are seeing & resulted in a poorly functioning (non functional really) server. When the upgrade was allowed to complete, ie until the point I saw "Remove CD & reboot", the upgraded server functioned perfectly without errors.

I think it is easy (but wrong) to assume the upgrade has completed when there is over 15 minutes of no apparent activity & the screen is saying "100% upgraded".

Yes the repos will be wrong & you will find that out using the advice in the FAQ, but also a lot of packages will be wrong or incompletely or incorrectly installed, & that's why you are now getting errors.

Are you absolutely certain that during the CD upgrade, you received the message "Remove CD & reboot", which would have been 5 - 30 minutes after seeing the "100% upgraded" message ?

Is it possible that you saw the message "100% upgraded", assumed it was complete, & then you rebooted using power switch or Ctrl Alt Del keys.
This is really a critical point to be sure about, or you will end up wasting a lot of time chasing numerous problems & errors, all of which are caused by the incomplete/interrupted upgrade.

[rafa0786]

Hi

Yes i am 100% sure it took about 10-15 min, the cd was ejected by the machine as well.

[janet]

rafa0786

OK that's fine.
You can easily check which repos are enabled in server manager, software installer panel.
The only repos that should be enabled (ie status=enabled) are those 6 shown in the wiki link.
You should enable any of those 6 that are not enabled. At this late stage of updating from sme 7 to sme8, you probably have base & updates repos disabled.

Once all 6 are correctly enabled then do
yum update

If any packages are installed or upgraded then follow with
signal-event post-upgrade
signal-event reboot

[CharlieBrady]

At this stage fixing the repos is unlikely to be sufficient to fix the problem. The symptoms indicate a mix between SME7 and SME8 packages being installed. That is almost certainly due to an interrupted/incomplete yum transaction, which has left some duplicate packages installed - both SME7 and SME8 versions of rpms. The SME7 rpms can be identified by 'el4' or 'sme7' appearing in their names in the output of 'rpm -qa'. They need to be removed.

The way to fix this is to start the system in single user mode, identify the duplicate rpms, and remove the older (sme7) ones using "rpm -e". Then reboot, and reconfigure, and the system should run OK.

A simpler solution is to restore from a good backup.