We will be performing an unscheduled firmware upgrade of the wireless infrastructure to address a bug in the current release which causes spontaneous reboots of wireless controllers and APs. These reboots have been causing service outages in various areas.
[Read more…] about 11:30PM 11/17 Emergency Wi-Fi Maintenance
Wireless Network Maintenance, 10/24/17
At 9PM on Tues. 10/24 we will be patching our wireless infrastructure in response to the WPA KRACK vulnerabilities. Wireless networks will be unavailable for 10-20 minutes while the infrastructure reboots.
[Read more…] about Wireless Network Maintenance, 10/24/17
Statement Regarding WPA KRACK Vulnerabilities
Earlier this week, several large implementation flaws were found in the Wi-Fi Protected Access (WPA) protocol which is used for encryption of wi-fi networks. The vast majority of wi-fi devices in the world are affected.
For general information about the vulnerability, please see:
https://www.krackattacks.com/
10 Common Vulnerability and Exposure (CVE) alerts were issued in response to this vulnerability. According to Cisco, 9 of these CVEs are exploited on the wi-fi client side and the remaining 1 is exploited on the infrastructure/AP side.
For the 9 client-side CVEs, users will need to update their client drivers and software to close the vulnerability. Many operating systems and hardware manufacturers have already made patches available. **Failure to patch your wireless devices may result in others being able to intercept and read any data you transmit over the wireless network.**
For the 1 infrastructure-side CVE, we currently expect Cisco to release a patch on Monday the 23rd. Previously the ETA was today, and we intended to patch over the weekend. Another IRIS News item will be posted once we have downloaded the patch file and scheduled a maintenance window.
EECS Server Room Power Backups Offline
We have been without backup power in our main server room, 290 Soda, since Monday due to an issue with our UPS (Uninterruptable Power Supply). All services are currently online, but in the event of a loss of power to 290 Soda all services would be offline for an extended period of time.
The UPS was due for preventative maintenance to have several capacitor banks replaced. This was conducted on Monday by a technician sent by the manufacturer, Schneider Electric. Unfortunately the technician made an error which resulted in failure of the UPS’ rectifier, preventing the UPS from being brought back online.
The vendor is working to locate a replacement rectifier, but the component is currently back ordered. When the replacement is received, a new technician will be dispatched with intent to restore the UPS as soon as possible.
This post will be updated if there are any major changes to the situation or when the repairs are completed.
[Read more…] about EECS Server Room Power Backups Offline
Wireless Network Frequency Issues, 10/20/17
All wi-fi networks except those at BWRC are currently impaired due to a wireless controller issue. The repair would be more disruptive than the issue so we will not take action until tonight.
The problem is that the vast majority of our 5GHz wireless radios are currently using the same wi-fi channel, causing a large amount of co-channel interference which is greatly degrading wireless performance.
Last night we made changes to the configured bandwidth of the 5GHz channels in an attempt to increase general reliability. Unfortunately we did not notice until well after the fact that the wireless controller’s dynamic channel assignment (DCA) algorithm was malfunctioning; this resulted in all or most radios defaulting to channel 36 and being unable to change.
Correction of the issue will require a full outage of the wireless network, which is why with user feedback we have decided to wait until after the business day to make any changes.
[Read more…] about Wireless Network Frequency Issues, 10/20/17
SDH, Blum, Calvin Network Outage 10/15/17
* Sutardja Dai Hall wired and wireless
* Blum Hall wireless
* Calvin Lab wireless
Network service was restored before 11AM. Please email help@eecs if you are still having trouble.
Resolved as of 2017-10-15 10:49:00
EECS-Secure Authentication Unavailable 10/11/17
The alert included a fix which we applied and service was restored by 13:35. During the outage, machines were unable to connect to EECS-Secure.
Information from the vendor about the bad update was posted at the following URL: http://community.arubanetworks.com/t5/Security/ClearPass-Policy-Server-is-down-after-the-recent-AV-AS-update/td-p/309698
Please contact help@eecs.berkeley.edu if you are still experiencing network problems.
Resolved as of 2017-10-11 13:35:00
Soda 5F South Network Issue
We are currently experiencing brief, intermittent losses of connectivity to Soda’s 5B network closet which serves the south side of the fifth floor.
The outages last approximately one minute and are spaced several hours apart. They appeared to be due to a fault in the 5B switch’s primary uplink to the BDF, so the primary link has been temporarily disabled in favor of the backup link.
As the issue is not currently severe we plan to investigate further on Monday. If things escalate, we will intervene sooner.
[Read more…] about Soda 5F South Network Issue
Network Border Maintenance, 9/18/17
Between 12 and 1PM on Monday 9/18, we will be migrating the department routing core’s uplinks to our new edge router. No interruption is expected but a brief outage is possible if we encounter a problem.
Please contact help@eecs.berkeley.edu with any questions.
[Read more…] about Network Border Maintenance, 9/18/17
BWRC Router Upgrade, 9/12/17
We will be upgrading the BWRC router at 8AM on Tuesday 9/12/17. An outage of approximately 10 minutes is expected.
The upgrade is intended to bring the router up to a vendor-supported software level, as well as to enable some new features.
Please contact help@eecs.berkeley.edu if you have any questions or concerns.
[Read more…] about BWRC Router Upgrade, 9/12/17