The Micros Server and cash registers were moved into a separate vlan on October 28, 2003 in an attempt to overcome system problems. The cash registers were having problems operating on-line in the congested Wheelock subnet. After some period of time, the cash registers would lock up, failing to process transactions.
The cash registers connection to the Micros Server seems much more stable now that they are in a separate vlan with a much smaller broadcast domain.
On October 17, 2003, the network in Trimble Hall was divided into two vlans. Vlan 116 contains all the Resnet ports while Vlan 100 contains the public ports located in the Forum and Seminar Rooms. The uplink from Trimble to the Core router was converted from a Layer 3 routed port to a Layer 2 Vlan Switched port. This configuration will be the prototype for the rest of the campus network.
The Cascade web interface was down for a several hours on the morning of October 8, 2003 following the regularly scheduled Information Services preventative maintenance period.
This outage was partially the result of some miscommunication within Information Services, and the improper shutdown of the Oracle Application Server on the effected system, Camano.
The backup server was brought on line around 11:00 am and remained up until late in the afternoon. The issues surrounding the failure are being addressed both within Information Services and with Oracle.
The production instance of the Oracle Application Server issue was resolved, and brought back on-line by the 5:00 pm.
Continue reading
On October 8, 2003 during Information Services normal preventative maintenance window, all systems with Solaris 8 and Solaris 9 received Sun’s recommended OS patches.
As a result of CERT Advisory CA-2003-26 and minor changes in the named daemon by ISC, we have upgraded both the named daemon on the internal and external DNS servers.
Apache, mod_fastcgi, mod_ssl, php, and openssl were all upgraded to the follwoing versions on the University’s main web server, www.ups.edu:
| Component | Version |
| apache | 1.3.28 |
| mod_fastcgi | 2.4.0 |
| mod_ssl | 2.8.15 |
| php | 4.3.3 |
| OpenSSL | 0.9.7c |
As a result of CERT Advisory CA-2003-26 the openssl has been upgrade on the University mail server, web server and dns servers. We are now working on rebuilding applications which use the effected libraries (e.g. httpd, sendmail, named).
For reasons that are not entirely clear, Ohio State has revoked the license on Prosepero version 2. The only information available is on their web site at http://bones.med.ohio-state.edu/prospero/.
This morning, the version of Prospero on the main web server was dropped to version 1.4.
Prospero 3 is on the way, and should be available in mid-October.
The sendmail configuration files on the University’s mail server have been upgraded to version 8.12.10. The early problem with the configuration files and the vacation program have been resolved. The vacation program has also been upgrade to prevent rev-lag.
The sendmail configuration files were also modified to use the currect os_type.
Problems were encountered with the sendmail 8.12.10 configuration file and the auto-responder (vacation) message. The configuration file has been rolled back while further investigation in being done.
The University mail server was upgraded from Sendmail version 8.12.9-20030924 to version 8.12.10. This was done as a more proactive step to maintain better version levels. The previous version was patched to address a OOB issue which was a temporary fix. The daemon and configuration files have been upgrade to the current release level of 8.12.10.
Sendmail on the University mail server was patched to prevent a buffer overflow condition as reported by CERT.
CERT Advisory CA-2003-25 Buffer Overflow in Sendmail
The patch used was taken from sendmail.org
A number of people in Resnet, mostly centered in the Langlow subnet, but not restricted to that area, had been reporting receiving invalid IP addresses, even though they had successfully registered their machines. The dhcpd.conf file had valid bootp entries for these machines as well. An astute student gave us a crucial piece of information – students that had registered after a certain time (unclear, but probably last week) were having the problem, while students resgistered before that time were working properly. This led me to believe that the dhcpd daemon was not running with the current version of its .conf file, and that a restart of the daemon would resolve the problem. When we attempted to restart the dhcpd process, it was found that the restart_dhcpd perl script was not running, whihc would explain the problem. Upon restarting the dhcpd, the problem was resolved.
HERMIONE had an unusually high CPU utilization (100%), and has for the past few days. This caused response times on Webmail become extrememly long, and the machine in general to become unresponsive. This was apparently caused by the presence of two FlexLM processes on the machine. Paul Monaghan gave me the history of this – he changed versions and apparently the old version did not get disabled.
Paul disabled the old version, and CPU utilitzation returned to more reasonable levels. Webmail response times lowered dramatically.
Measures taken against the Nachi worm, namely the blocking of TCP ports 135 and 139, caused the Best Lock and Video monitoring systems to fail. Apparently, theses systems use these ports to communicate with the Lenel Communication Server Process.
The blocks were lifted and the Communication Server was restarted, which restored the connection.