Monday around 10PM I need to reboot the host machines to load a new kernel. Ubuntu keeps kicking them out and most have had security issues so don’t want to leave the old ones in place any longer than necessary. There should be about ten minutes of hard downtime and then, particularly on the debian machines, may require additional reboots if they do not remount NFS partitions right which has been a problem with them recently.
I am going to take the new web server down tonight for about 20 minutes to image the machine now that I have it stable and operating well.
There have been 0 owncloud errors and 0 php errors in the last 24 hours. This machine has never run this clean. Ubuntu definitely runs much better than CentOS.
An additional plus, it takes approximately three seconds for a reboot as opposed to close to two minutes under CentOS.
The meeting was a bust, maybe half a dozen people verses about two dozen at the previous meeting. I did get some good ideas for Eskimo’s future from those few but I am disappointed that there wasn’t more participation.
I need your input, was the location bad, time bad, the more formal format bad, what kept you away this time?
What can I do to make the next meeting a success?
I did not save the configure options I used when I compiled the ftp server back in 2012 on the last incarnation of our ftp and web server. Consequently I had to guess at some of them.
One bad guess was –paranoid, which said, “Disable some questionable features”. It turns out two of those features was the ability to remove or overwrite a file.
This broke a number of weather stations and the ability for one person to update his website.
I have removed that option which seems to have resolved these problems.
I’ve enabled a bunch more functionality in ownCloud that could not be installed with the old web server owing to php module compatibility issues which are resolved on the new server. There is a really nice calendar manager now as well as a note pad and other productivity functions.
Under the Information pull down there is now an entry User Meetings that provides information on the next time and place (tomorrow at Razzi’s Pizza).
The new web server is online.
If you are using memcached is your friend or wp fragment cache in your WordPress installation, you may have to de-install and re-install these for your site to work properly.
Since these break your ability to login, you will need to go into shell, go to the root of your wordpress, then cd to wp-content, and then to plugins. Use rm -r to erase fragment cache or memcache. Then you will be able to login, delete them, and re-install and everything should be okay.
In the past I have recommended shellx to use for a development platform for websites hosted here as it was the same platform that the web server was based upon.
After the server change tonight, Ubuntu will be the preferred platform to use for development of your website here as the new web server will be based on Ubuntu.
If there are any tools that you need that aren’t presently on that machine, please let me know.
Friday evening I will be replacing the existing web server. The new server will utilize the same hardware infrastructure (i7-6700k at 4.2 Ghz) but will differ in software.
The operating system will change from CentOS 6.7 to Ubuntu 15.10.
MySQL and PHP will be native to the distribution instead of compiled separately which should eliminate some compatibility issues that are presently breaking redis which is used by ownCloud causing the loss of some ownCloud functionality.
The new mysql installation will utilize a native mysql driver which is more tightly integrated and this will improve response time for LAMP applications slightly. The current average response time is 100ms, this should shave a few off of that.
Initially display_errors and display_startup_errors will be turned on so I can check out and debug the new machine after it is turned up. Unfortunately, some debugging can not be done until it is online and serving traffic. In particular phpMyAdmin is not functioning yet but it may be because of the domain.
There will be some downtime involved because I have to dump the existing MySQL databases, copy them to the new server, and import them into the MySQL on that machine. Also I will need to change the IP address and the loading of virtual machine addresses because there can’t be more than one machine listening to a given IP address at a time.
I was able to resolve the problem with wu-ftpd so no ftp function will be lost in this upgrade.
This will make future upgrades less service interrupting because Ubuntu can be updated live where CentOS can not, at least not with applications or third party software installed.
It will make it easier to bring modern web applications online because the kernel and libraries will be current instead of 3-1/2 years old now as is the case with CentOS.
I will post to News (here) when the new machine is online so that if you notice any problems you can generate a ticket (Support->Tickets). If you generate a ticket please be sure to include the URL that is not being served correctly.
Our router at the co-location facility cabinet spontaneously rebooted at 3:10PM this afternoon. This resulted in an outage lasting approximately 2 minutes.