Hi,

Webalizer suddenly stopped working after 15 november on a few of our RAQ's
I changed the configuration so that webalizer was processing our logs before they are rotated in a script that's rotating the system logs, (/etc/logrotate.d/apache)
but two servers are working great and the rest not.
I tried to start webalizer manually, without the logrotate script. But it gives no errors, and is finished very quickly after I start it.
What can I do to debug my situation here ? Or do you know a solution ?

Thanks in advance.

We had the same problem with October 4. You can solve it by upgrading to the latest version of Webalizer. Just downlad the tar.gz, extract en replace the old webalizer executable by the new one, thats all.

yep but that was a bug which I solved with the patch for it.
If I upgrade to webalizer 2 it must be done the 1st of the next month, otherwise I will lose statistics. But my customers can not wait till then.

Can i just copy the new 2.0 webalizer.pl file over the old 1.3 ?
Or will it cause major problems ?

Webalizer isn't working at all at the moment and i can't upgrade
with the cobalt package, so this would be an easy and welcome solution !

I have the same problem. My webalizer is version 2.0.1 . Webalizer stop working since 9 November 2001. Can someone point out a solution for me? Thanks.

You'll get the same problem for any site that is over the disk quota.

I'd check that first.

No. There's still alot of Disk space. Everysite 's webalizer on the server stop working .

Have you try running it from the command line?

Originally posted by yesyes
Hi,

Webalizer suddenly stopped working after 15 november on a few of our RAQ's
I changed the configuration so that webalizer was processing our logs before they are rotated in a script that's rotating the system logs, (/etc/logrotate.d/apache)
but two servers are working great and the rest not.
I tried to start webalizer manually, without the logrotate script. But it gives no errors, and is finished very quickly after I start it.
What can I do to debug my situation here ? Or do you know a solution ?

Thanks in advance.

I renamed logrotate to zlogrotate in the cron.daily folder.

This way the logrotate is run after webalizer is completed.
No changes to webalizer itself was made and this works great.

when I run webalizer manually on a log file, it says it can't read the log file. the path is good, I gave the file chmod 777....
when I open the log file with more , the syntax is good.
I really think this is a bug.
so still no statistics on most of our servers.

NewonNet: I run ./webalizer.pl

Right now, the problem isn't solved yet. If there's nothing i can do, i have to uninstall and install webazlier again. Any article that teaches us how to do this?

in /etc/cron.daily directory, do you have these files?

logrotate
webilzer.pl

and in /usr/bin

webalizer2.pl
webazolver

Yup. I have those 4 things.

When you run webalizer.pl, did you get any errors?

The program should run for a while before returning if you have large log files.

sorry to interrupt agian, but i have this webalizer problem for quite some time now.

If i should remove these mentioned webalizer files by hand, can i consider it uninstalled ? or do i need something else for that ?

Thanks !

To uninstall any packages, you should run unintall script in /var/lib/cobalt/uninstallers/ directory.

Thank you,
i tried that but it doesn't work.
Upgrading from the control panel also doesn't work.

So i thought i remove everything by hand and start all over..

We have been having the same problem on our 2 Raq's... the 1st couple of sites we put on work fine with /~stats, but the more recent ones don't work at all.... I keep getting emails from cron saying :

Warning: Truncating oversized referrer field
Warning: Truncating oversized referrer field
Warning: Truncating oversized referrer field
Warning: Truncating oversized referrer field
Warning: Truncating oversized referrer field
Error: Can't change directory to /home/sites/ns2.bennettandco.co.uk/users/stats/web
Error: Can't change directory to /home/sites/www.heartsmart.uk.com/users/stats/web

Could this be something to do with it???

Nope don't think so.
This truncating stuff, are just log entries that are to long to process if i remember well.
I a bit weird i think.
You made a site for your nameserver (ns2?), it obviously did'nt get any hits.
The whole directory seem a bit strange to me, its searching for a directory /web in stats ???
Mind you, i don't know a awful lot of this stuff, although i'm learning quick with all these bugs lately :)

The answer to my question i got from Bradford L. Barrett, i think the author of webalizer.
Maybe someone else can use it to their benefit, it helped me.
He got it from a user and forwarded it to me.

------------------------------------------------------
1) keep webalizer 1.3 installed on the system, and if you uninstalled it go
back and install the
same package again.


2) Telnet into your box and su for root


3) cd to /home and makdir webalizer or anything else you like


4) cd to that directory and at the prompt type :
wget ftp://ftp.mrunix.net/pub/webalizer/webalizer-2.01-06-linuxelf-bin.tgz
now you need to untar the file you just downloaded so do this
tar zxvf webalizer-2.01-06-linuxelf-bin.tgz
now the file is untared now at the prompt type cd webalizer-2.01-06-linuxelf


5)now you need to copy the file webalizer into /usr/bin/webalizer
cp webalizer /usr/bin/webalizer


6)now you need to get pnglib file and here is how to get it:
wget
ftp://speakeasy.rpmfind.net/linux/redhat/6.2/en/os/i386/RedHat/RPMS/libpng-1.0.5-3.i386.rpm


7)now just do this:
rpm -i libpng-1.0.5-3.i386.rpm


and that's all there is to it if you encounter any problems check to see if
you typed everything
right. If everything went ok and you want to test it right away you can do
this:
./etc/lib/webalizer.pl
this will take some time to finish , after that you can go to your stats page
and you should see all
your stats before and after the 4th of october.
-----------------------------------------------

After i did all this, i started webalizer, and got a 99 error.
To fix that i had to manually (for 135 sites :( ) remove "webalizer.current" from all the stats directories for each site.

Hope someone can use all this.

... with the logrotation of the stupid RaQ and not WebAlyzer. This happenes when the logfiles are big.

Solution SSH/Telnet in as SU and type:

split_logs web </home/log/httpd/access

You will notice the next day that the cron jobs resumed activity and WebAlyzer works again.

Originally posted by ShareFile
... with the logrotation of the stupid RaQ and not WebAlyzer. This happenes when the logfiles are big.

Solution SSH/Telnet in as SU and type:

split_logs web </home/log/httpd/access

You will notice the next day that the cron jobs resumed activity and WebAlyzer works again.

wats that command u wrote sharefile
can u share some useful information as to wat does it do and how does it help
to solve the problem:rolleyes:

The RaQ has a big logfile for ALL activity. If everything goes as planned ("as planned" on the RAQ haha good joke!) a daily cron job breaks it down to smaller logfiles per virtual site.

These smaller log files are being processed by WebAlyzer then.

Now if you have some traffic and the file gets really big, the logfile rotation blows. As the file grows even bigger a compression job compresses the logfile and starts a new one. At thi spoint it is too late.

So, when you see that WebAlyzer does not show stats for 1-2 days, it is most likely that above scenario occured.
So what we do in this case is to rotate the logfiles manually by telneting/sshing into the RaQ, login as SuperUser (su -) and type the command "split_logs web </home/log/httpd/access" and pres enter.

Now take a coffee break, have a shower, go out, have kids and hopefully after all that the job will have finished :D

Do we need to force the system to zip the access file after this procedure?

My log files is about 10MB per day. I probably need to rotate the logs twice per day.

How big the access file must get before the system is forced to zip the file?

Originally posted by NewonNet
Do we need to force the system to zip the access file after this procedure?

No no, just do what I told you and inform me about the results. If it doesn't help then we will look further and if it DOES help we will se to the rest... one step at a time ;)

thanx for all the help sharefile
i did wat u told and got my stats to work

but will i have to do that split_logs thing manually every day

Originally posted by SixthSense
thanx for all the help sharefile
i did wat u told and got my stats to work

but will i have to do that split_logs thing manually every day

1) Glad it worked for you (I struggled HOURS with Cobalt Support until we (them and me) discovered this "yet another cobalt glitch"...

2) No you don't, just when you see that it is stuck again. It's like a toilet... you only call the plumber when it spills lol

The solution work wonders!

Thanks for the helpful info.

... seems like a lot of ppl. had this problem... must be coming right after the "email aliases issue" :rolleyes:

Isn't there a config we can do with logrotate on the file size issue so that it does not break?

I'll have to look into that.

Originally posted by ShareFile
The RaQ has a big logfile for ALL activity. If everything goes as planned ("as planned" on the RAQ haha good joke!) a daily cron job breaks it down to smaller logfiles per virtual site.

These smaller log files are being processed by WebAlyzer then.

Now if you have some traffic and the file gets really big, the logfile rotation blows. As the file grows even bigger a compression job compresses the logfile and starts a new one. At thi spoint it is too late.

So, when you see that WebAlyzer does not show stats for 1-2 days, it is most likely that above scenario occured.
So what we do in this case is to rotate the logfiles manually by telneting/sshing into the RaQ, login as SuperUser (su -) and type the command "split_logs web </home/log/httpd/access" and pres enter.

Now take a coffee break, have a shower, go out, have kids and hopefully after all that the job will have finished :D

Hummm... I'm trying to do this and here are the results:

[root /etc]# cd /
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]# su
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]#

Any help please?

Originally posted by gertiebeth


Hummm... I'm trying to do this and here are the results:

[root /etc]# cd /
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]# su
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]# split_logs web </home/log/httpd/access
sh: split_logs: command not found
[root /]#

Any help please?

You have a RaQ 3 maybe?

Originally posted by ShareFile


You have a RaQ 3 maybe?

No, it's a 4...

I've crashed 4 days in a row, all at the same time. Here is a sample output of mu kernel log right before the crash:

Jan 4 18:37:59 www kernel: RPC: task of released request still queued!
Jan 4 18:37:59 www kernel: RPC: (task is on xprt_pending)
Jan 4 18:37:59 www kernel: portmap: RPC call returned error 111
Jan 4 18:37:59 www kernel: RPC: task of released request still queued!
Jan 4 18:37:59 www kernel: RPC: (task is on xprt_pending)
Jan 4 18:37:59 www kernel: lockd_up: makesock failed, error=-111
Jan 4 18:37:59 www kernel: portmap: RPC call returned error 111
Jan 4 18:37:59 www kernel: RPC: task of released request still queued!
Jan 4 18:37:59 www kernel: RPC: (task is on xprt_pending)
Jan 5 06:07:00 www kernel: VM: do_try_to_free_pages failed for httpd...
Jan 5 06:12:32 www kernel: Unable to load interpreter /lib/ld-linux.so.2
Jan 5 06:13:29 www last message repeated 2 times
Jan 5 06:18:48 www last message repeated 17 times
Jan 5 06:19:53 www last message repeated 20 times
Jan 5 06:20:55 www last message repeated 19 times
Jan 5 06:22:27 www last message repeated 23 times
Jan 5 06:28:19 www last message repeated 61 times
Jan 5 06:29:25 www last message repeated 5 times
Jan 5 06:30:48 www last message repeated 11 times
Jan 5 06:37:35 www last message repeated 12 times
Jan 5 06:38:48 www last message repeated 13 times
Jan 5 06:39:54 www last message repeated 12 times
Jan 5 06:40:55 www last message repeated 63 times
Jan 5 08:46:27 www kernel: klogd 1.3-3, log source = /proc/kmsg started.
Jan 5 08:46:27 www kernel: Inspecting /boot/System.map
Jan 5 08:46:28 www kernel: Loaded 7617 symbols from /boot/System.map.
Jan 5 08:46:28 www kernel: Symbols match kernel version 2.2.16.
Jan 5 08:46:28 www kernel: Loaded 124 symbols from 2 modules.
Jan 5 08:46:28 www kernel: Linux version 2.2.16C28_III (root@sun.cobalt.com) (gcc version egcs-2.91.66 199903$
Jan 5 08:46:28 www kernel: Ignoring bogus EBDA pointer 3FFF000
Jan 5 08:46:28 www kernel: Detected 448219 kHz processor.

After doing some research here (I LOVE this place!) I came to the conclusion that I needed to split my log files because they were too large. But now I can't figure out how to do it! :bawling:

You appear to have a serious problem which has either been caused by a malfunction (probably hardware) or from something you installed.

The only advice at this point from me is to do a OS-Restore from the Restore CD and to contact Cobalt support.

Originally posted by ShareFile
You appear to have a serious problem which has either been caused by a malfunction (probably hardware) or from something you installed.

The only advice at this point from me is to do a OS-Restore from the Restore CD and to contact Cobalt support.

Thanks Sharefile. I'll go back to crying now. :bawling:

Originally posted by ShareFile
You appear to have a serious problem which has either been caused by a malfunction (probably hardware) or from something you installed.

The only advice at this point from me is to do a OS-Restore from the Restore CD and to contact Cobalt support.

OK, I just downloaded the log file and set it to support via my admin server manager. I have tried this twice before (for other problems) but I've never heard back from them...

Originally posted by gertiebeth


OK, I just downloaded the log file and set it to support via my admin server manager. I have tried this twice before (for other problems) but I've never heard back from them...

Call Cobalt Europe Support in the Netherlands/Europe (they speak english) an dask for Jaques Zambli and / or David...
You get patched through faster when you know a name there :D

Never mind.. my mistake