Difference between revisions of "WebMaster"
m (→Webmaster Status: Updated status)
|Line 25:||Line 25:|
== Webmaster Status ==
== Webmaster Status ==
Revision as of 08:20, 22 December 2005
The Eclipse WebMaster deals with the servers and software that runs the eclipse.org site. We don't code Java (or at least we try not to) and we don't maintain a lot of the eclipse.org website content.
Who is the WebMaster
Denis Roy: IT Manager, SysAdmin
Matt Ward: SysAdmin
What does the WebMaster do?
So you don't do Java or HTML, what do you do? - We manage and monitor the eclipse.org servers and infrastructure - We respond to requests directed at firstname.lastname@example.org - We take care of our Committers - We take care of the main downloads pages - We code some PHP tools for the website - We respond and fix bugs assigned to the Community product in Bugzilla
Contacting the WebMaster
Please peruse to the Webmaster FAQ before contacting the WebMaster.
Dec 22, 2005 Things have calmed down since Phoenix was released, and we we able to finish our MBOs. I'm working on the Build environment (bug 117052) and Matt is putting together some slick new newsgroups and mailing list pages.
Merry Christmas, and a Jolly Good New Year!
Dec 2, 2005 We are currently flooded with requests. Expect response delays. We apologize for the inconvenience.
Nov 30, 2005 Phoenix is released! We still have many pages to update and move over to the new style, but it's a big start. Here are some issues we're working on:
- Move the news server over to a smaller box, and configure the 550 as a build server
- Get Bugzilla to use the shadowdb correctly (broken in 2.20)
- Keep on truckin' with Phoenix - move the news, mail and search engine pages to the new template
- Update dev.eclipse.org pages to look fresh and new
- Committer tools enhancements
- Improve Foundation internal workflow
It's non-stop action behind the webmaster desk.
Nov 9, 2005 Busy busy busy. Matt and I are busy ripping our hair out with the following:
- Sandbox OTRS.
- Spend a week fixing Community bugs
- Work on Phoenix
- Installing and configuring a web analysis tool
- Committer tools enhancements
- Sandbox bugzilla 2.20 for upgrade, then upgrade once tests are done
- Improve Foundation internal workflow
- Test and update the Eclipse.org disaster-recovery plan
As usual, committer requests to email@example.com are our top-priority, as are ensuring the eclipse.org infrastructure rns smoothly.
Sept 20, 2005 With Matt now on board and up to speed, things are back to normal. Not a moment too soon. Matt is currently moving our in-house foundation.eclipse.org stuff from a PC to real servers (HP Proliant DL 360's) and I'm working on the Foundation's database application.
August 8, 2005 I've finally caught up to the backlog of WebMaster requests. My inbox is empty. Let this be a lesson to me: never go on holidays again.
August 2, 2005 I'm currently working on a backlog of webmaster requests. I was on holidays last week, and Sujay left the Foundation on Monday July 25, so there's a week's worth of unprocessed requests that I'm tending to.
As of this writing (14:39) I'm up to 2:00pm July 25, with 90 e-mails left. I hope to have everything caught up by Thursday, Aug. 4th. I apologize for the delays.
July 17, 2005
- This is just a test to see if the automagic checkout mechanism still works after the big migration.
July 5, 2005 What a busy time for us. After spending countless hours making sure we survive Release 3.1 with the least bruises possible, here's what we're up to:
- preparing the IBM p5 550 servers for the migration. Two of these powerful servers will assume backend duties at eclipse.org. We're going to transfer data from our current backend server, test the migration, test it again, then start all over until we can get it right in the least amount of time. Scheduled date for the transition is Sunday, July 17 2005 at 6:00am Eastern.
- working on a web-based database application for the Foundation's use.
- we're also going to start customizing PHPNuke, the CMS we chose for Project Phoenix
June 15, 2005
- We're currently in the process of: assigning top-priority to WebMaster requests for projects that are following the 3.1 release schedule
- configuring the IBM p5 550 servers for a fault-tolerant backend. This will increase available disk space and speed up database requests. The current backend server will become "node4" and will assist the three other nodes in servicing front-line requests
- creating a Disaster Recovery plan for eclipse.org
- setting up project Phoenix
May 24, 2005 There has been some discussion on Ryan's blog regarding PlanetEclipse noise and blogging guidelines. Right now, there are no real guidelines per se; we encourage subscribers to supply the URL to their Eclipse-related feed, and because of the low number of subscribers, PlanetEclipse is relatively easy to read.
However, as the number of subscribers increases, so will the number of articles, and there is concern that noise will eventually become a problem.
So here are the questions we need answers to in drafting posting guidelines:
- What is your definition of noise?
- Should we allow/encourage non-Eclipse articles (movie reviews, gadgets, misc rants)?
- Should we be subscribing anyone to the PlanetEclipse aggregator, or should we perform some screening?
- What happens if we subscribe someone and they post more noise than actual content?
Please post your comments in this bugzilla bug and read Ryan's blog for more comments. Help us make PlanetEclipse a useful resource for the entire community!
Apr 28, 2005 I recently setup PlanetEclipse, a news feed aggregator for the Eclipse community. Here is what I will be working on for the following weeks:
- A plugin signing process. Anyone have experience with keytool/jarsigner? The man pages are driving me nuts.
- Eclipse.org infrastructure support tools: internal workflow, work request system, web-based tools for committers, etc...
- Finish setting up the IBM hardware. I can't wait to see these beasts in action.
I am glad to say that the backlog of webmaster requests is all cleared up, so my response time to your requests will be back to its normal pre-migration snappiness.
Apr 12, 2005 The final crumbs left from the migration are being dealt with. Aside form the QoS rules which took a while to figure out, everything went pretty smoothly. As it stands, we now have a scalable and fault-tolerant eclipse.org: I can easily add new servers to handle requests, and eclipse.org can suffer the total loss of one (or more) servers without having much impact on the overall functionality. Although the backend is currently not redundant, it will be in the months to come. I'm currently working on the backlog of webmaster requests, as well as finishing up the last migration crumbs and implementing new features and tools for committers.
Mar 26, 2005 dev.eclipse.org is moved over to the new hardware. What a rush! There are still a few things left to shuffle around, but the big piece is done. Now that it's moved over, there will be the usual kinks and issues to solve then I will have more time to answer firstname.lastname@example.org requests. Thanks again for all your patience while the migration took place.
Mar 19, 2005 I just finished moving the Eclipse web site and newsgroups over to our new hardware The transition went quite painlessly because of massive pre-migration testing. So far the new hardware isn't breaking a sweat handling downloads, bugzilla, the website and the newsgroups. The real test will be next weekend, when the CVS repositories get moved over. Prepare for some good speed increases to dev.eclipse.org!
Many thanks to HP, IBM and Intel for the fantastic hardware.. it rocks!
Mar 15, 2005 I'm smack dab in the middle of moving everything over to the new hardware. eclipse.org ig going to rock once this id done. However, it's keeping me quite busy, and my response time to the email@example.com inbox requests is seriously affected by this. I am doing my best to respond to urgent manners in a timely fashion.
Feb 10, 2005 I'm currently in the process of preparing new hardware infrastructure. My response time to the firstname.lastname@example.org inbox will be seriously affected by this, but I'll do my best to respond to urgent manners in a timely fashion.
Jan 28, 2005 OUTAGE ALERT! Eclipse's bugzilla service will be down for about 2 hours while I upgrade bugzilla to the latest 2.18 version. I will begin the upgrade Wednesday, Feb. 2/2005 at 4:00am EST and should be complete by 6:00am EST.
Other things tying me down: - Preparing the new hardware infrastructure
Jan 13, 2005 Happy New Year! I've just updated the Webmaster FAQ. Besides tending to webmaster requests, I'm also: - Preparing the new downloads interface and "find a mirror" script for community feedback. - Preparing the new hardware infrastructure
Dec 22, 2004 Tomorrow, Dec. 23, I will be out of the office setting up the new backend servers for Eclipse.org. I will only respond to email@example.com inquiries that are urgent. Thanks for your understanding. Happy holidays!
Dec 15, 2004 This week I'm working on the following: - Getting project websites out of ViewCVS and onto the new repository. Read Dec. 7 for more details. - Getting project download pages to use the "find a mirror" script - Working on a simple front-end to query download stats for the projects using the "find a mirror" script - Planning the upcoming hardware migration
There are currently no firstname.lastname@example.org request backlogs.
Dec 7, 2004 GOOD NEWS! Projects hosting web content using ViewCVS on dev.eclipse.org can be moved over to a new repository that is automatically transferred to www.eclipse.org/your_project. This eliminates ViewCVS, long URLs, speeds up page display, enables authors to create dynamic PHP content and opens the door to using a MySQL backend. If you'd like your project's web site to be moved over to the new repository, drop me a line at email@example.com and I'll get the gears rolling.
I'm also working on the following: - co-ordinating with IBM, HP, Intel and Magma for the new hardware shippment - Moving Hyades over to TPTP. - restoring service to our stats server. It seems to have been "out of it" since Oct. 18. - Dealing with the backlog of webmaster requests
Although I'll do my best to respond to all queries in a timely fashion, please expect some delays when sending requests to firstname.lastname@example.org for the next couple of weeks.
November 30, 2004 I am currently testing the following: - New downloads UI - CVS-enabled webpage authoring for projects
November 15, 2004 This week I am finishing up the recent migration and tending to the webmaster Inbox. I am also continuing my work on the mirrors and downloads architecture.
November 11, 2004 I'm am planning this saturday's migration to our new location, and I want everything to go as smoothly as possible. Also, I will be meeting with an infrastructure engineer on Thursday and Friday to start preliminary planning of the new infrastructure. For these reasons, you may experience delays in receiving a response from me. I apologize in advance.
November 4, 2004 I'm currently tending to my inbox requests..
October 26, 2004 I'm currently tending to my inbox requests. I've received lots of feedback about how difficult it is to download files from projects other than Eclipse Platform because these files are not mirrored, so I'm currently working on this infrastructure issue.
October 20, 2004 I'm currently tending to the backlog of requests to the webmaster inbox. As a result of planning the infrastructure migration, several requests were left unanswered. New mirror requests will be done this week as well.
Dec 18, 2005
news.eclipse.org will be down for a few minutes Sunday, Dec. 18 2005 between 6:00am and 8:00am EDT while I move the news server to a temporary box. This is to accomplish what's described in bug 117052. Long story short, the P5 server I want to use for the build environment currently hosts the news server, and will continue to do so later, but it's not configured properly (only one CPU is configured out of 4, disks in software raid config, etc). Back in March 2005 news.eclipse.org was transferred to this ill-configured P5 box temporarily because of the server migration deadline I had to meet. news.eclipse.org will be interrupted again after Christmas to move it back to the freshly configured P5, where it will live happily ever after. Actual downtime was 13 minutes
Nov 20, 2005
Bugzilla will be upgraded to the latest version in our outage window (6:00am Eastern). Expected downtime will be 3 hours. The upgrade process performs heavy database-intensive updates for over 2 hours. According to Bugzilla, a "large bugzilla installation" is >10,000 bugs. We have about 110,000. Actual downtime was 27 minutes
October 29-31, 2004
Phase II of the eclipse.org migration to the Eclipse Foundation will take place the weekend of Oct. 29-30-31. During this time period, the followiing services will be intermittently unavailable.
main eclipse website online help bugzilla bug tracking and reporting stellation.eclipse.org
December 15, 2003
Our internet provider will be conducting maintenance from 00:00-02:00 EST Tuesday Dec 16, 2003 to resolve some issues wit h one of our internet lines. During this time period, the followiing services will be intermittently unavailable.
downloads via ftp and http rsync for eclipse mirrors online help bugzilla bug tracking and reporting stellation.eclipse.org
December 14, 2003 All eclipse.org servers will intermittently unavailable from 9-10 am EST for routine software maintenance.
November 16, 2003 The download.eclipse.org servers will intermittently unavailable from 9-11am for software maintenance. The bugs.eclipse.org server will be unavailable for 30 minutes during this timeframe for minor hardware maintenance.
October 5, 2003 The dev.eclipse.org and www.eclipse.org server will be down Sunday Sunday October 5, 2003 8-9am EDT for minor hardware and software maintenance.
September 13, 2003 The dev.eclipse.org server was down Saturday, September 13 8:00-8:30am EDT for minor hardware maintenance.
September 3, 2003 The download.eclipse.org, bugs.eclipse.org and stellation.eclipse.org servers was unavailable for a 15 minutes during the 3am until 6am EDT time period September 3, 2003 to allow our carrier to conduct maintenance.