This is the WordPress ChristmasPress theme slightly modified.
If you’d like a WordPress website setup either as your website, or as part of your website, it is free to asking for any Eskimo North customer. Just e-mail support, let us know your login, where you want the WordPress site setup, if you have an existing MySQL database here we will need the password, if not give us a password to use to set one up.
I will set it up for you totally free of charge.
If you’re not an Eskimo North customer, with the purchase of any shell or web hosting package, I’ll be glad to do this for you.
Shellx is restored and I’m going to leave it alone until early Saturday morning so as not to cause further unintentional outages.
There are a number of errors in the start-up scripts which I need to correct but I need to reboot multiple times to test in order to do so.
In the process of trying to clean up some start-up issues on Shellx, I’ve broken it in ways that I can’t recover from except to go to a backup. It will be down for about 20 minutes to recover from a backup image.
I apologize for the interruptions in web, mail, and shellx service today. This was necessary because the upgrade to CentOS 6.5 broke the NFSv4 clients badly.
NFS is networked file system, it’s how common files like your home directory and mail spool are available on multiple machines.
NFSv4 doesn’t pass numerical user ID’s over the wire like previous version of NFS did. Instead, it passes names which are mapped by idmapd, which up until recently worked.
Unfortunately, the NFS version 4 client in CentOS 6.5 is sometimes passing a numerical UID over the wire instead of a name and then when idmapd tries to look it up, it fails, resulting in the file ownership being mapped to nobody.
Reading the CentOS and RedHat forums, it’s clear that we’re not the only people having this problem, it’s also clear the developers don’t yet have a clue what is causing it, so a fix may be awhile.
I apologize for the somewhat slower response this is going to cause in the meantime but random user ID mapping is not a workable situation.
That portion of our system maintenance which is service affecting has been completed.
Mammoth Networks, which provides DSL back haul for our DSL customers using CenturyLink circuits, will still be rebooting equipment at 2AM PST which will cause an interruption of approximately 15 minutes duration in DSL connectivity.
Mammoth Networks, the company we use for DSL back haul for CenturyLink circuits, will be rebooting a router at 2AM Saturday morning. Connectivity will be affected for approximately 15 minutes.
We will be performing maintenance work early Saturday morning between just after midnight and about 2AM that will affect the main file server, shellx, mail, and ftp for about 20-25 minutes each which involves rebooting to make updates from Centos 6.4 to Centos 6.5 active, and imaging those machines so that if recovery is necessary, it will be to current software.
The main file server will affect all services and will be first, just after midnight, then the other services individually will be taken down to be imaged and rebooted.
At Xeno’s request, I’ve installed Ruby 2.0.0 on Shellx. However, because there were 128 things dependent upon the existing ruby, rather than overwriting the system version, I’ve installed 2.0.0 in /usr/local, the executable is in /usr/local/bin. If you want the newer version, either specify the full path or make sure that /usr/local/bin is before /usr/bin in your $PATH variable.
I removed a group that I didn’t realize was still in use and broke mail. The group has been put back in place.
This morning some website CGIs and potentially other active content refused to execute. This was caused by basically the same problem that affected the mail servers, idmapd not being properly configured and started. In the case of ftp it was configured but not enabled.
For reasons that escape me, somehow nfsv4 was able to usually resolve id’s even with idmapd not running prior to the upgrade the weekend before last, but since that upgrade it is now working as documented and requires this daemon to run properly.
It is now enabled on the webserver. This problem is resolved.