PALS Outages
(This page is obsolete since the service was discontinued July 31, 2006.)
Here are some suggestions of what to do if the AuthPals dll times out (result = 4) for all 4 server addresses.
- Our Monitoring — we usually know about all problems before you'd notice
- WhatsUp Pages and Logs — where you can look yourself (jump)
- Ping — first thing to try if no outage is logged
- Test Web Pages — next thing to try
- Contact Us — last thing to do
Our Monitoring
Note that the service is critical for the operation of our labs, and that we have monitoring software that repeatedly check both PALS servers and send messages to our cell phones whenever they don't reply. It is highly likely that a PALS outage is known by CLC staff within a minute or two. Note also that there are two completely separate servers, and they are on machines with two network adapters, each on separate networks. Also, the primary server is on a 2-node cluster, and the service will fail over to the other node in the event of a hardware failure.
However, we did have one occasion in 2002 when both servers were not responding (nothing is perfect!) That was bizarre, and we made some code changes that should circumvent that problem in the future.
WhatsUp Pages and Logs
As noted above, we run software that monitors the Pals servers and reports outages to us by email and SMS messages to our cell phones. This is custom software that we are actually phasing out in favor of a commercial package called WhatsUp Professional.
A special WhatsUp page for Pals is here. You can check the logs via a web page and see if there was an outage at a time in the past when your users couldn't log in. But note that network outages or problems may be logged as "misses" when the service was not really down because the WhatsUp server is not on the same network as any of the Pals servers.
The main WhatsUp page may also be useful for you to quick see if there are campus networking problems.
Ping
The first thing to do is ping the pals server addresses. Ping the following:
- pals.win.psu.edu -- primary server
- palsb.win.psu.edu -- backup server
- palsc.win.psu.edu -- = primary server
- palsd.win.psu.edu -- = backup server
If none of the pings reply, then there is probably a network outage. To confirm this, ping:
- 128.118.155.1 -- network gateway for primary server
- 128.118.155.129 -- network gateway for backup server
We may soon spread the network connections amongst two other networks.
If neither of these answer the ping, then either our networks are down or yours is. In any case, the problem is not PALS. We suggest further diagnosis to eliminate your own network as a problem (e.g., can you get to other things outside your subnet) and then calling the TNS Network Operations Center at (814) 865-4662.
If some of the Pals address reply to ping, then the AuthPals dll ought to be working, but perhaps slower than usual. I'm not sure what to say here; go to the next step anyway.
Test Web Pages
If all the pings work, then that:
- doesn't mean that both Pals servers are working, or working well.
- doesn't entirely eliminate network problems; there have been cases in the past where ping worked fine but routers weren't passing UDP packets (that was fun).
Regardless, the next thing is to try our page that uses AuthPals dll, https://clc.its.psu.edu/AuthPals.aspx. You don't even have to use a valid userid and password. If that works, then the PALS service is ok.
Contact Us
If that page fails to get a reply from a Pals server, then please contact the CLC Hotline to be sure they know something is wrong.
This site maintained by the Classroom and Lab Computing group of Information Technology Services.
Suggestions and comments about this web site: CLC Webmasters; Other contacts here.
This page was last modified: 1/23/2007 9:26:28 AM.