cPanel 11.24.4 Enters RELEASE Today!
Server Admins (me included) the world over woke up today to find that a major upgrade to the cPanel/WHM control panel had been rolled out. Within the release are a multitude of upgrades to features within cPanel along with some major optimization to increase the performance speed and reduce overheads.
Sounds great so far? Yeah, well - This upgrade hasn’t exactly gone too smoothly for some folks.
Having put the first coffee on to brew I turned on the screens in my office to do the first scan of emails received overnight from my servers, and noticed one repeating theme from a particular box telling me (at regular 5 minute intervals) that FTP had failed. Hmmm, OK, a quick visit to WHM for that box, scroll to the bottom of the menu (Restart Services), hit “FTP server” and “OK”, and I’ll be heading back to the kitchen for that all-important first cup of coffee in no time, and then back to investigate the logs in a few minutes … Gotta get that coffee first!
WRONG!
For some inexplicable reason WHM isn’t playing ball. I know that I’ve entered the right password (It’s just a paste from my password manager) so why am I not seeing WHM load up? Instead I’m getting a 404 error, with the message:
The server was not able to find the document (./scripts3/initial_setup_wizard1) you requested.
Please check the url and try again. You might also want to report this error to your webhost.
Huh?
I glance across to another screen that’s showing my live monitoring and FTP is definitely showing green for that server. WTF? I happen to have a login to an account on that server saved in my FTP clent so I hit it quick and, behold, FTP is fine on the server - I logged in without a problem.
OK, decision time - I decide to go get that coffee (and a cigarette) and come straight back to this.
Safely armed with injections to hand of caffeine and nicotine I glance at some other emails and notice the one from cPanel announcing the upgrade release. “Cool” I think “I’ll check that out in a bit after I find out what’s wrong with this dang server”.
Glancing left to another screen I notice the amber flashing item in the toolbar which tells me there’s been new messages in the IRC channel we run between ourselves (my support guys and me). I open that up to find a few items about failed webmail logins, customers cannot access their cPanel logins, etc., etc.
Mostly, my over-night guys have fixed all the issues and our customers are all fairly content, but there still remains this issue of FTP and WHM on one server. In my book 2 + 2 generally equals 4, so I follow the link in the cPanel announcement email to see what’s changed.
To cut a long story short, the update roll-out hasn’t gone entirely like it should, and a number of issues have surfaced. Rather than post a blow-by-blow account of how the rest of today unfolded, I’ll paste below what we’ve found and how we fixed it, in the hope that it will be of use to someone else out there …..
1) The cPanel update servers have taken a battering so, if you have a box that’s slow, or is only getting a slow connection to the cPanel update servers, TailWatch has a habit of restarting processes the update stopped (before it’s ready for them to restart).
This causes the update to stall, so stop TailWatch first with killall -TERM tailwatchd, then run killall upcp to stop the failed update, then /scripts/upcp –force to fire it off again.
2) There’s LOTS of changes in this release, so it’s worth spending some time running through all the options in WHM to make sure everything is as you want it …. and secure!
For example, I just took a look at the new FTP configuration options to find 2 settings set to YES by default:
Allow Anonymous Uploads This is set to YES and right next to it is a warning about how it affects server security … doh!
Allow Logins with Root Password This is also set to YES! So, someone guesses/sniffs/cracks a root password and then can merrily FTP into any account on the box and upload malicious scripts, even if you’ve secured SSH, changed the port, etc, etc. NOT NICE!
3) Message just in from cPanel:
Please ensure to read the following message carefully as there is important information about MySQL and Automated process killers in relation to the cPanel 11.24 upgrade. MySQL Root Password ——————————— With cPanel 11.24 now available, it is imperative to ensure that you have set your MySQL root password before upgrading to the latest version of cPanel. If you have previously skipped this step, you can set it now using the “MySQL Root Password” function in WebHost Manager. Failure to set a MySQL root password may cause database corruption on systems running MySQL 4.1 with InnoDB tables. Please set your MySQL root password as soon as possible to avoid any issues. Machines without a MySQL root password set allow access to any database by any user so it is imperative that a password is set as soon as possible. If you have moved /root —————————— If you have set root’s home directory to something other than /root, you will need to copy .my.cnf from root’s home directory to /root/my.cnf. In the future, this process will be automated. Automated process killers ———————————— If you have a system in place that automatically kills processes taking up a lot of CPU time, you should disable them before running the cPanel update as the size of the update has triggered many of these systems. These process killers will kill updates or related process in the middle of the process and cause issues. If you have a failed update, you can force a re-update by running /scripts/upcp –force
If you have already updated to cPanel 11.24 and experienced this problem please open a ticket at https://tickets.cpanel.net/submit/Thank You,
Eric Gregory
cPanel, Inc.
Fortunately, that one didn’t get us.
4) DNS Clusters: If you have more than one VPS/server and are running a DNS cluster in WHM, then check it after this update.
Normally, when you access Configure Cluster from WHM you would see a list of hostnames that are configured within the cluster. If WHM is now showing just the IP addresses instead (in the left hand column), then the trust relationship between the servers is possibly only working in one direction. This can screw up the parity of DNS records on all servers in the cluster very quickly.
A quick fix we’ve discovered is to renew the trust relationship in BOTH directions - In other words, install the remote access key for one server on the other, and then repeat it from the other server back to the first (previously the trust relationship would be established both ways by just installing a key from one server to the other).
Once you have the trust relationship installed both ways, WHM will then display the hostname rather than the IP and all should be well.
If you have users/customers who are likely to have created/deleted domains or sub-domains to their accounts since the cPanel update was applied, you might also want to run “Synchronize all zones to this server only” on all servers (one at a time) to ensure that all your records match.
So far so good - All my boxes are running and we don’t have any major disasters (unlike some), but we HAVE disabled automatic updates from cPanel on all boxes apart from one test box that’s allowed to break … Once bitten, twice shy!
And today’s Phrase of the Day, as seen in numerous PuTTy windows during the course of today … “Performing sanity check“


