v14.1 Release - Bugfixes, Maintenance and More

Fantastic news!

Thank you for all the effort you put into this.

During initialisation the HDD indicator was very busy. But once the firstboot scripts completed it settled down. I logged into Redmine within my web browser and it all seemed to be running fine to me. It wasn't super fast but considering that Redmine is written on Ruby on Rails and the low resources allocated, it ran as good as I would expect. With no activity (other than the web page still open in my browser) the vHDD indicator was occasionally "blipping" but otherwise was mostly off. Actually the network indicator was "blipping" as much, if not more than the vHDD indicator.

Although for my test to be conclusive I probably should update VirtualBox and install from ISO as you did (and install initial sec-updates too by the sounds).

Assuming that it is given the same virtual hardware, theoretically once the ISO is installed the VM should be identical. Essentially that's how we build the OVA - by (scripted) install of the ISO to a vHDD and then packaging it. Out of interest is there a reason why you used the ISO and not the OVA (pre-built VM)?

Did you use a sparse or a pre-allocated vHDD? Do you have plenty of free (real) harddrive space? Do you have a lot of other i/o operations going on? Do you have plenty of free RAM (i.e. you system isn't paging/swapping too much)?

The only thing I can imagine is that if you used a sparse disk and you don't have a lot of free space and/or there are a lot of other i/o operations going on (including paging/swapping) then the VM may have cached a lot of it's writes (to RAM - which may in turn have then been paged/swapped) and a whole vicious cycle is going on. Then add on top of that extending the sparse disk on the fly; that will often cause a flood of i/o.

FWIW my system is i7 laptop; 4GB RAM; 256GB SSD with 4GB swap allocated. I'm running Debian (as mentioned above) with Gnome3, Thunderbird running and about 20 Chrome tabs open. With the 14.1 Redmine VM running my system has 170MB RAM free & 2GB swap free and everything is running pretty sweet...

I suggest that you leave it running for a little while and see what happens. And/or try rebooting it. And/or try to free some other system resources and see if that helps it settle down...

Have you had a go of the Webmin "Let's Encrypt" functionality? TBH I'd be really interested to hear about the use of it. I didn't actually get a chance to test it out but read about it in the Webmin changelog.

FWIW we have big plans for integrating Let's Encrypt into v14.2. The plan at this stage is to include the option of pre-seeding Let's Encrypt cert generation (so it will be automated within the Hub at launch or can be pre-seeded for all the other builds). We are not quite sure exactly how users will configure it post-install/post-firstboot (if not pre-seeded) but we don't really want to add extra interactive firstboot scripts. Actually we'd like to reduce them - there already too many questions at firstboot IMO. So it is likely that we will extend the confconsole to include additional configuration options, but we'll have to see how that goes...

As per always, your input and feedback are warmly welcome! :)

Another excellent TKLX team effort. The Let's Encrypt and Java-less Webmin are certainly timely and much appreciated enhancemenets.

Congrats on corralling all the moving parts Jeremy :-)

Thanks for your kind words! :)

Actually something that I've thought of is that there may be some appliance specific config changes that you may like to make on your appliances (which are a part of v14.1). To check out whether that may be the case, you'll want to have a look at the bugs we resolved for v14.1. Go to GitHub and looking at the v14.1 "milestone" issues. Then select your appliance name from the "Label" drop down (see image for details).

As much as possible I try to add all the appliances that a particular bug affects, however sometimes I will only add "core" if it affects all appliances, or "lamp" if it affects all lamp based appliances. So I suggest that you also check both "core" and probably "lamp" too. Unfortunately thought GitHub doesn't support OR operations on tags/labels, so you'll need to do the searches individually. Hopefully I haven't confused you too much... If you need any clarification (either on what you should check for your specific appliances, or how to implement a specific change/bug fix on your appliance) please do not hesitate to ask...

TBH I think that the TurnKey website needs a whole lot more than just a m.turnkey.org! It's in some serious need of some love I reckon! We are hoping to do a whole website refresh at some point in the not-too-distant future. I think it has been nearly 6 years since the last one...!

Hi, thnx it really helped, because I don´t like bugs haha (nobady does :P) keep up the good work!

p.s maybe the turnkey needs a mobile version like m.turnkey.org :P

apt-get update
apt-get upgrade
apt-get dist-upgrade

With regard to the updated upstream software; unfortunately there is no easy way to update that without risking data loss. So really you'll want to do that manually anyway. FWIW with most of the appliances; even if you migrate your data with TKLBAM (from v14.0 to v14.1) then your backup will still include the old version of the upstream software; so you will still need to update that yourself...

Can I upgrade my v14.0 to v14.1 with the command line or something? or must I reinstall my dedicated server?

As of v14.0 our config is slightly different WRT default Debian (Jessie) Apache (2.4), but not significantly different.

The default global Apache cert location is defined in the ssl_mod config (/etc/apache2/mods-available/ssl.conf). So you could add your new cert to /etc/ssl/private and adjust the default global Apache SSL conf to point to it. Alternatively you can define certs on a site by site basis by adding the 'SSLCertificateFile /path/to/cert.pem' directive within the 443 server declaration in your current site file (default in LAMP is /etc/apache2/sites-available/000-default.conf). Either will require an Apache restart (service apache2 restart).

To use the cert with Webmin and Webshell you'll need to also adjust the stunnel config (/etc/stunnel/stunnel.conf). If your cert is causing issues then double check that your pem looks something like this:

-----BEGIN CERTIFICATE-----

    ... cert here ...

-----END CERTIFICATE-----
-----BEGIN PRIVATE KEY-----

    ... private key here ...

-----END PRIVATE KEY-----
-----BEGIN DH PARAMETERS-----

    ... Diffie Hellman parameters here ...

-----END DH PARAMETERS-----

If your doesn't look like that then you should be able to make it look like that using the current pem, crt & key files that you currently have. TBH I have limited experience with "proper" (i.e. not self-signed) SSL certs so am unable to double check any of my understandings.

Another thing to be wary of is that if you created (or edited) the pem file in Windows Notepad it may have mangled the line endings? I recommend that you use something like NotePad++ for Windows editing of any text files that are intended for use by/in Linux.

Also one of our community members has provided a script (turnkey-make-ssl-cert) which is now in all TurnKey appliances by default. TBH I have only used it with it's default configuration (to create self-signed certs) but by my understanding it can be used to generate other certs.

I hope at least some of that helps...

FWIW as a general rule, posting on our forums is the best way to ensure a (more) timely response. I monitor those regularly and there is a nice list of unanswered forum posts that I use when responding on the forums. I usually closely monitor blog posts for the few weeks after they're published. I also get notifications about comments on blog posts too, but they can often get buried in my inbox. As there isn't a nice list anywhere of unanswered blog post comments, the only time I usually see "late comments" is when I clean up spam or otherwise find myself on a blog post.

Anyway, it sounds like the permissions are too locked down to allow upgrade of SuiteCRM in our appliance?! Changing the ownership of ./ (probably /var/www/suitecrm) to www-data should resolve that.

If you need a hand with something, please start a new thread, rather than posting on a really old blog post.

Please note that starting a new thread, requires you to be logged in as a (free) registered website user.

In your new thread, please provide as much info as possible regarding the appliance (name and version) you are having issues with, where it's running (i.e. what platform) and the nature of the problem you are experiencing. Please include explicit error messages you are seeing and/or screenshots to assist you to clearly describe the problem. For example, how I could recreate the problem that you are experiencing.