Home › Blogs › proco's blog

AppServer Installation Errors

Thu, 2008-08-07 11:50 — proco
0
No votes yet

We are in the process of configuring our production environment for parallel deployment. We are trying to create our second AppServer in advance of getting the load balancer.

Yesterday we tried to do the install of LP-4.0.0.0-linux.iso on a clean server running Red Hat AS3 Update 9. The install created large backup file (backup-luminis-4.0.0.0.zip) of about 7 Gig before ultimately failing with the following error

----------
2008-08-06 11:14 ERROR unexpected exception occurred
Traceback (most recent call last):
File "/usr/cpsource/salina/installer/_freeze/linux.freeze/bin/main.py", line 322, in _install
loop.install( inst, ifile, propsfile )
File "/usr/cpsource/salina/installer/_freeze/linux.freeze/bin/loop.py", line 194, in install
bak.close()
File "/usr/cpsource/salina/installer/_freeze/linux.freeze/bin/backup.py", line 72, in close
self.zip.close()
File "/usr/local/lib/python2.4/zipfile.py", line 503, in close
OverflowError: long int too large to convert to int
2008-08-06 11:14 ERROR [E000] An unexpected error has occurred.
long int too large to convert to int
----------

We opened a ticket with Sungard and they said this error is caused by logs being too large. But, being a clean system, we don't have very many logs and none of them are very large.

We are wondering if anyone has experienced this and/or has suggestions on how to get around it. This second server is named b-app2-lumprod and our first one (which is fine) is b-app1-lumprod.

I've included our installation config file below. Is there possibly something in there that we have missed? Thanks!

Mike Procopio
Web Services Developer
Seton Hall University
procopmi@shu.edu

------------------- b-app2-lumprod-install.conf-----------
# What to install
tiers = portal

# Where to install
cp.root = /usr/luminis

# How to connect to LDAP running on the Resource box
resource.directory.manager.dn = cn=Directory Manager
resource.directory.manager.password = ********
resource.host = b-app1-lumprod.shu.edu
resource.ldap.port = 389
resource.ldap.suffix = shu

# Properties required for the Resource box
admin.id = lumadmin
admin.password = ********
shared.directory = /usr/luminisshared
grouptools.chat.port = 9257

# Properties required on Unix
os.group.id = luminis
os.user.id = lumadmin

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Fri, 2008-08-08 10:57 — proco

Found problem, encountered new one

Well, we found out what why the install was creating such an obscenely large backup file. The installer was trying to back up our Java directory! In the directory is a symbolic link, 'latest', which points to the current directory. So, not only was the installer trying to back up Java, but it was trying to do so *recursively*!

We've now encountered a new problem. We have two servers: b-app1-lumprod & b-app2-lumprod. b-app1 is the server hosting the resources and b-app2 is just an AppServer. When we try to go to b-app2, it gets redirected to b-app1 and all our links point to b-app1.

We talked with Sungard and they said this is acceptable behavior, but that seems awfully strange to us. When we put our load balancer on top of this, won't all the traffic going to b-app2 from the balancer get redirected to b-app1?

We suspect there was a problem with our hosts file when we did the install. So we're going back to a clean OS and reinstalling again. We'll also be adding "local.host=b-app2-lumprod.shu.edu" to our install configuration file.