Difference between revisions of "Wikispooks:Site Status"
m (more) |
m (add link to heartbleed info) |
||
| (One intermediate revision by the same user not shown) | |||
| Line 2: | Line 2: | ||
The two values for system, status are NORMAL/UNSTABLE. Be careful not to introduce whitespace here, as it is affects page layout elsewhere | The two values for system, status are NORMAL/UNSTABLE. Be careful not to introduce whitespace here, as it is affects page layout elsewhere | ||
| − | -->{{Site status| | + | -->{{Site status|UNSTABLE}}<noinclude> |
<br/> | <br/> | ||
'''If you hit a problem that you think is to do with the site software, report it [[{{TALKPAGENAME}}|here]].''' | '''If you hit a problem that you think is to do with the site software, report it [[{{TALKPAGENAME}}|here]].''' | ||
---- | ---- | ||
<br/> | <br/> | ||
| + | |||
| + | ==24 April 2014== | ||
| + | The site was affected by the [http://en.wikipedia.org/wiki/Heartbleed Heartbleed vulnerability] until 1045GMT today. The fix has caused an access issue with the server security module which in being worked on. Expect the odd server re-boot until further notice. --[[User:Peter|Peter P]] ([[User talk:Peter|talk]]) 11:02, 24 April 2014 (IST) | ||
==Site outage for at least 8 hours on Saturday morning 22 Feb 14== | ==Site outage for at least 8 hours on Saturday morning 22 Feb 14== | ||
Revision as of 10:04, 24 April 2014
If you hit a problem that you think is to do with the site software, report it here.
Contents
24 April 2014
The site was affected by the Heartbleed vulnerability until 1045GMT today. The fix has caused an access issue with the server security module which in being worked on. Expect the odd server re-boot until further notice. --Peter P (talk) 11:02, 24 April 2014 (IST)
Site outage for at least 8 hours on Saturday morning 22 Feb 14
This is reported to me by the hosting company as having been a 'domain registrar status issue' - which makes me uneasy. I've asked for further info. --Peter P (talk) 12:21, 22 February 2014 (GMT)
- Further info: from Host support "... essentially it's a glitch in the control panel that crops up form time to time where a security feature in the control panel kicks in in error and disables a domain. There's nothing you could have done about it on your end. Apologies for the inconvenience".
Upgrading SMW and Semantic Forms
Upgrading semantic software withe the php cache and Scribunto disabled. This will also require a data rebuid. Site will remain operational but usual cautions apply - should be back to normal by 12:00 today. --Peter P (talk) 08:27, 12 February 2014 (GMT)
Scribunto/SMW issue
Recent refresh of SMW data threw up lots of errors (@150 out of 9,000 items refreshed). The error was generated by a Scribunto Lua-Engine script called by the SMW refresh script in the course of its looping execution. I intend to switch off Scribunto and run the SMW refresh again to help understand what is happening here. Site notice should be up for maybe 2 hours or so from this message --Peter P (talk) 07:26, 9 February 2014 (GMT)
- Refresh done - issues with 14 records (lots less than before) but still don't know what's causing them. --Peter P (talk) 08:14, 9 February 2014 (GMT)
Move to new server
Status of the move and any issues arising.
To-Do list of issues arising from the upgrade' - please add to it if you spot anything.
Saturday 18 January 2014
Move complete and most issues resolved You should notice major server response improvements. There's still some server tweeking to do, notably to Apache mod_security and php cache tuning. These will require occasional Apache restarts so best to keep copying edit box contents to clipboards before saving. But I'm taking a bit of a break from intense techie-tweeking for a few days --Peter P (talk) 08:34, 18 January 2014 (GMT)
Friday 17 January 2014
The site will be put offline from 0700 GMT for about 1 hour while a number of maintenance routines are run. --Peter P (talk) 19:31, 16 January 2014 (GMT)
- All data refreshed. There is minor issue arising from my having left the $wgServer directive set to the new server IP address (rather then the hostname for DNS reasons while testing) for an hour or so after going live with it. It has caused 9 'record ID's' out of about 11,000 to have 'invalid server index entries'. I don't know for sure which actual pages that involves right now but it will be those edited very soon after going live - and could be the cause of the Image problems on the pages Patrick edited. Also this page since I was quick to post a progress report on it. It's not serious and will be resolved over the next few days. Meantime please report any other issues. --Peter P (talk) 09:12, 17 January 2014 (GMT)
- Panic over pages disappearing from the Main Page boxes resolved. After running SMW_refreshDate twice to confirm the index ID's of problem pages, I ran it once more on just the SMW Type and Property pages (Completes very quickly). It seems that breaks their links to containing pages so as the jobs queue progress the pages disappeared from the Main Page boxes. It is necessary to run SMW_refreshData on both Types/properties AND regular pages even though doing types and properties first negates the need for two complete passes. BIG learning curve this :-)) --Peter P (talk) 13:12, 17 January 2014 (GMT)
Thursday 16 January 2014
The site is now running on the new Cloud-based VM server
There are likely to be teething issues. Please advise here and I'll see to them asap. We have access to the old server for another month or so. It is not open to the world but all software and files remain in place until the contract expires - Insurance :-)).
Note: I will be restarting the Web server several times through today to apply and test configuration changes as they are made, so please continue to exercise caution per the red notice above for as long as it and the main page banner warning are in place --Peter P (talk) 08:10, 16 January 2014 (GMT)
- Major issues resolved so site banner warning removed. MAY need to apply minor changes to web server config but your browser cache should cope if you get a 'server reset' message. Piwik, Cron-job backups setup, and the anon-ul glitch should be done today too. --Peter P (talk) 12:45, 16 January 2014 (GMT)
| ||||||
|
