My next steps are getting those OSL boxes serving requests. At the moment, 
Docker Swarm seems like the least-terrible way to do so, but that means making  
the applications' configurations more cloud-friendly so configs and secrets can 
be managed safely. I think I've got a plan for that, just trying to iterate in 
my head over how maintainable I can get it...

The need for 2-3 rather large storage devices for raw report data remains. 
Didn't Ruth say she had an in with a cloud vendor via TPF as well? Though... 
the "backup" DB server has been non-functional for a while... And we still have 
a contact at ServerCentral (Joel Berger) so that hosting isn't as tenuous as 
ByteMark was (past tense now... unfortunate...)

And the Rexfiles for deploying things still need some work. I fixed what I 
could from installing on my Debian Trixie VM, but ideally they could deploy to 
Rocky and we'd have it deploy to Docker Swarm (or something cloudy) instead.



Doug Bell
[email protected]



> On Nov 23, 2025, at 5:39 PM, Scott Baker <[email protected]> wrote:
> 
> Doug!
> 
> You are a superstar my friend. I can confirm that log.txt is working now, and 
> so is the API to fetch test results. Magpie is happy again, so that mean's 
> Slaven's stuff is probably also working. What are the next steps? How can we 
> help?
> 
> I/we/the Perl community really appreciate your hard work getting CPT mostly 
> functional again.
> 
> -- Scottchiefbaker
> 
> On 11/23/2025 11:48 AM, Scott Baker wrote:
>> Don't worry about "E-mailed test reports". I've been brainstorming how we 
>> can do that with Magpie and the API it already has. I'll handle that piece.
>> 
>> Others are welcome to step in as they can.
>> 
>> -- Scottchiefbaker
>> 
>> On 11/23/2025 7:00 AM, Doug Bell wrote:
>>> Summary:
>>> 
>>> * Our main server died
>>> * We're running on a small VM for now
>>> * We will scale to more VMs today
>>> * Some things are broken and cannot be restored without contributed labor
>>> 
>>> ---
>>> 
>>> Yesterday sometime before 10:30am US/Central, the primary CPAN Testers 
>>> application server died: One of the main disks failed. Since this machine 
>>> is hosted by ByteMark, and since we have no contact with them anymore, 
>>> there is no chance at restoration. As of 19:00 US/Central, I was able to 
>>> restore operations on a medium-sized VM I'm renting from NetActuate[1]. It 
>>> is less than half the capacity of the bare-metal machine that died, so 
>>> we're still not back to where we were.
>>> 
>>> Scott Baker has procured some VMs from the Open Source Lab at Oregon State 
>>> University[2]. I am going to scale the app to those systems as well in hope 
>>> that the three VMs will handle the load we generate. My guess for that is 
>>> "probably".
>>> 
>>> But, with ByteMark's servers now completely out, there are parts of the 
>>> site that simply do not function. 
>>> 
>>> * Our CPAN/BackPAN mirror is gone. I'm working on requirements to see how 
>>> this can be fixed.
>>> * E-mailed test reports have been offline for a while, but now is gone 
>>> until a replacement can be made.
>>> * Large sections of the primary site are still unimplemented
>>> 
>>> If anybody wants to make a large impact for a little work on some 
>>> Mojolicious sites, we can get things working. Until then, I'm doing what I 
>>> can with what I have.
>>> 
>>> 
>>> 
>>> Doug Bell
>>> [email protected] <mailto:[email protected]>
>>> 
>>> 
>>> 
>>> 
>>> 1: https://www.netactuate.com/
>>> 2: http://osuosl.org/
>>> 

Reply via email to