|
Home > Archive > Netware Webserver > October 2005 > Apache2 won't start from clustered volume
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Apache2 won't start from clustered volume
|
|
| jschick@nospam.org 2005-10-12, 5:55 pm |
| Gerben,
I'm assuming your volume resource is separate from your apache resource,
and online on the same node before you try to launch the apache resource.
In theory this should work fine; same with using protected memory space (as
long as it's not Netstorage, for one). You may try adding the server name
or IP address after the -f in your load line just in case.
I approached it a different way, by keeping Apache on the SYS volume of my
nodes and creating individual .conf files for each instance we run
(iManager, Netstorage, and a web-enabled phone book). The ones that have
data on clustered volumes point to those volumes via their respective .conf
files.
Justin
jschick at aafp dot org
| |
| Gerben Welter 2005-10-13, 5:51 pm |
| Hi Justin.
I have apache and the htdocs on the same resource and is offcourse
online. I even start apache from that resource.
I tried your suggestion, but adding the servername or ip address
doesn't make a difference.
Maybe this is the same issue I've seen with clustering groupwise. The
POA needed to login to the cluster resource, although it was running
on the same server.
I guess I have to use a workaround like only putting the htdocs on the
cluster and manually replicating any changes in the httpd.conf to the
other nodes.
Gerben.
On Wed, 12 Oct 2005 20:12:58 GMT, jschick@nospam.org wrote:
>Gerben,
>
>I'm assuming your volume resource is separate from your apache resource,
>and online on the same node before you try to launch the apache resource.
>
>In theory this should work fine; same with using protected memory space (as
>long as it's not Netstorage, for one). You may try adding the server name
>or IP address after the -f in your load line just in case.
>
>I approached it a different way, by keeping Apache on the SYS volume of my
>nodes and creating individual .conf files for each instance we run
>(iManager, Netstorage, and a web-enabled phone book). The ones that have
>data on clustered volumes point to those volumes via their respective .conf
>files.
>
>Justin
>jschick at aafp dot org
| |
| Gerben Welter 2005-10-13, 5:51 pm |
| I just let apache start from sys and configured it with the
apacheadmin. I made a vhost that has its document root on the webvol:
so I know apache can browse the clustervolume. But it still doesn't
work! Logs show it can't find it. If I move that document root to SYS:
it works. What's happening here?!
Gerben.
On Wed, 12 Oct 2005 20:12:58 GMT, jschick@nospam.org wrote:
>Gerben,
>
>I'm assuming your volume resource is separate from your apache resource,
>and online on the same node before you try to launch the apache resource.
>
>In theory this should work fine; same with using protected memory space (as
>long as it's not Netstorage, for one). You may try adding the server name
>or IP address after the -f in your load line just in case.
>
>I approached it a different way, by keeping Apache on the SYS volume of my
>nodes and creating individual .conf files for each instance we run
>(iManager, Netstorage, and a web-enabled phone book). The ones that have
>data on clustered volumes point to those volumes via their respective .conf
>files.
>
>Justin
>jschick at aafp dot org
| |
| jschick@nospam.org 2005-10-13, 5:51 pm |
| > I have apache and the htdocs on the same resource and is offcourse
> online. I even start apache from that resource.
Could it just be a timing issue?
If I'm reading that right your clustered volume is being mounted and Apache
is being ran from the same cluster resource. If so, it could be that the
script is just executing too fast and Apache just isn't able to see the
volume yet.
To test I would have the volume resource online and then manually run
Apache from the console. If that works you should just have to add a delay
in the load script between the volume mount and the call to Apache, or
create a separate resource just for your apache instance.
Justin
jschick at aafp dot org
| |
| Gerben Welter 2005-10-24, 9:14 am |
|
No, it's not a timing issue. I'd leave the volume mounted and manually
restart apache. For the mean time we're going to run apache on one
node and gather more info. If we don't solve this soon, we're probably
going to open an incident with Novell.
Gerben.
On Thu, 13 Oct 2005 19:22:20 GMT, jschick@nospam.org wrote:
>
>Could it just be a timing issue?
>
>If I'm reading that right your clustered volume is being mounted and Apache
>is being ran from the same cluster resource. If so, it could be that the
>script is just executing too fast and Apache just isn't able to see the
>volume yet.
>
>To test I would have the volume resource online and then manually run
>Apache from the console. If that works you should just have to add a delay
>in the load script between the volume mount and the call to Apache, or
>create a separate resource just for your apache instance.
>
>Justin
>jschick at aafp dot org
| |
| jschick@nospam.org 2005-10-24, 9:14 am |
| >
> No, it's not a timing issue. I'd leave the volume mounted and manually
> restart apache. For the mean time we're going to run apache on one
> node and gather more info. If we don't solve this soon, we're probably
> going to open an incident with Novell.
>
> Gerben.
Good luck!
Justin
|
|
|
|
|