04.30.08
Posted in WSO, services, website at 12:16 am by Ben
After a couple reports of problems with the site recently, we’ve done some updates to fix a few bugs. You may have seen a little Service Temporarily Unavailable, Server Error, or Maintenance message this evening while we did the updates.
- Facebook photo uploads are fixed. (For the past 2 weeks, JPEG uploads didn’t work)
- Announcements should now properly assign posts to the section you tell it to use. (there was a bug where a new announcement would land in the Announcements section no matter where you posted it if you had to re-login when you clicked the Submit button.)
- “Random” 500 Internal Server Errors - They shouldn’t show up so much (any???) now. Please let us know if they do. The problems came from moving session storage to the database and storing massive amounts of session data (e.g. giant facebook query results, whole user records, etc.). Some tweaks to Facebook have simplified things to remove these problems, which were really leftover from the live search days and manifested by moving session to the db.
- Broken links at the bottom of the home page are corrected. Double check them.
- Rides are now sorted intelligently.
Permalink
Comments off
04.10.08
Posted in WSO, systems, services at 9:26 am by Ben
Now that the WSO machines have moved into our new rack, we’ll be doing some major software updates on our web server and some reconfiguration of user services this Friday into Saturday, starting after 6pm on Friday and continuing on Saturday. During that time, a variety of services will be unavailable. We don’t have an exact timeline, but here’s the shortlist of what’s going down and what’s not.
- Mail forwarding and mailing lists will remain available the entire time.
- The main WSO website will be down almost the entire time.
- User and organization websites and ssh logins will be down for a shorter time.
UPDATE: We’re back on the web just after 8pm Saturday. There will be a short bit of downtime later this evening, but it should be, er, short. Please let us know at webmaster at wso.williams.edu if something that used to work is broken…
Permalink
Comments off
04.07.08
Posted in WSO, systems, services, facilities at 10:51 pm by Ben
All WSO services will be down starting at 4pm and hopefully ending by 5:30 or 6:00 on Wednesday April 9th. During this time, all WSO services, including the website, email forwarding, email lists, user logins, etc., will be unavailable. All email sent to wso.williams.edu addresses during this time will be queued at other servers and sent through later in the day. (We’ll try to minimize downtime of the mail server too…)
WSO has been outfitted with a shiny new four-post rack complete with KVM, etc. thanks to Gabriel McHale at OIT. We’ll be moving into it on Wednesday and getting all our machines nice and tidied up. (The current setup doesn’t quite exude an aura of organization…)
Update:
We finished the move in a little under 3 hours. Planning and organizing a rack layout took longer than expected, but it’s really nice to see the results. We’ll post pictures somewhere sometime soon if we get around to taking them.
Permalink
Comments off
02.07.08
Posted in WSO, website at 5:15 pm by Ben
So now that the new home page has been out a few days, things seem to be quieting down. We’re slowly adding or tweaking features on the page in response to people’s opinions.
New today:
- a little more color (purple section headings. Not much, I know)
- “helpful” links, clustered at the bottom of the page.
Originally, getting the links off was a first-class goal of the design, but it looks like putting the links at the bottom seems to fit into the scheme of the rest of the page without introduce too much more clutter. The top of the page represents links to commonly used extra-WSO services, followed by WSO services. The middle is recent content, and the bottom is the mass of links to useful but hard to find pages. Putting the links in a block instead of a sidebar also means that
- we don’t sacrifice any width in the layout.
- you can see them all at the same time. No more scrolling up and down trying to find the right category and then the right link. You can do all that with your eyes.
To do:
- Icons for services.
- alternate color scheme (user chooses)?
Permalink
Comments off
Posted in WSO, website at 5:05 pm by Ben
That server error you might’ve seen was the database filling up with session info (that tracks info like what your user name is when you’re logged in). Due to an oversight when switching from file-based to database-based sessions, we neglected to empty them periodically, resulting in the DB “filling up.” That’s been fixed, so hopefully we won’t be bitten by this again.
Now, once you’re logged in, your session expires after an hour (of inactivity, I think.)
Permalink
Comments off
02.06.08
Posted in WSO at 8:31 am by Ben
The 30-40 minute power outage this morning knocked out the firewall again. I checked connections with Todd Gould in NetSys this morning and it turns out that half of our machines were only on generator power, not the magical in-floor UPS. Everything is hooked up to UPS power now, so this should not happen again in the future.
Permalink
Comments off
02.02.08
Posted in WSO at 1:17 pm by Ben
About to take the site down (hopefully for a brief period) to deploy new home page, factrak, rides.
Permalink
Comments off
02.01.08
Posted in WSO, systems, accounts at 8:54 am by Ben
Ursula, the user server, seems to be among the missing this morning. I’m headed over to see what’s up soon.
Update: A reboot fixed things. All is back to normal. (I still need to take a spin through the logs to see what happened.)
Permalink
Comments off
12.28.07
Posted in systems, services at 11:56 am by Ben
Yesterday at 4:00pm Jesup lost power. Although the UPSes were supposed to cover our systems and most of our machines seemed to stay up, the firewall went into a zombie state and wouldn’t keep any connections. Thanks to Pete C. at OIT, the firewall has been rebooted (c.a. 11:15 this morning) and we’re back in business. We’ll be checking our power connections with Pete to make sure we’re plugged into the right places. During the period from about 4pm yesterday to 11:15ish am today, no services were available. Any mail sent to wso addresses during this time was queue by other servers and will all have trickled through within a few hours.
Permalink
11.03.07
Posted in systems, services, website at 6:23 am by Ben
Nancy’s network handling seems to have gone slightly flaky this morning. Apache was not receiving any incoming traffic on port 80 despite running. A reboot fixed things at about 7:20am.
Permalink
« Previous entries ·