Post migration/upgrade proxy errors and delays

[Cristy Danford]

Hello all,

I'm working on a local server migration to Rhel 8 and concurrent upgrade of ASpace (v341 to v411), as well as upgrading Solr and MariaDB to compatible versions - via traditional deployment (VM, no Docker). We have a custom PUI plugin which required some updates, but aside from this, there were very few issues. The Solr dashboard loads near immediately, indexing was a breeze, and logging set to debug indicates no issues I can identify in archivesspace.out. ASpace appears ready to go within a couple of minutes, which is in line with our production server.

Unfortunately, we (long before my time) deviated from the recommended practice of using a subdomain and instead utilize a prefix (http://aspace.myarchive.org/staff and http://aspace.myarchive.org/public). Via the new setup, both the public and staff pages take 20 minutes to load, and deliver 502 proxy errors in the browser and the httpd logs in the interim. 

I've never worked with ProxyPass or Proxy URLs before, and TCP/IP is not a strong suit of mine. I've been troubleshooting this alongside our IT Dept, various LLMs, etc., and nothing seems to make a difference. I disabled all plugins. I've uncommented all default URLs in config.rb and changed them to 127.0.0.1, rather than localhost. I've tried every version of localhost/127.0.0.1/domain in the frontend and public_proxy_url lines in config.rb, as well as via the ProxyPass and ProxyPass Reverse directives in the Apache config, although perhaps a bit haphazardly, as I'm not clear on best practices. I've researched thread dumps and explored port types/labels.

I would love any suggestions the group has to offer, and my fingers are crossed that someone out there has experienced this! I'd be delighted to be embarrassed by some small error at this point. 

Thanks for any assistance,

Cristy Danford
Library and Academic Systems Technologist
Barnard College

[Cristy Danford]

To be clear, this is just upon start up/restart of ASpace. After 20 minutes of proxy errors/delay, both the public and staff interfaces load normally from there on out. Sorry about that!

Cristy

[Blake Carver]

Did you check the ArchivesSpace logs during that time, was it running or was it just starting up?

---

From: archivesspac...@lyrasislists.org <archivesspac...@lyrasislists.org> on behalf of Cristy Danford <cdan...@barnard.edu>
Sent: Thursday, February 5, 2026 4:56 PM
To: Archivesspace_Users_Group <archivesspac...@lyrasislists.org>
Cc: Cristy Danford <cdan...@barnard.edu>
Subject: [ArchivesSpace Users Group] Re: Post migration/upgrade proxy errors and delays

 

You don't often get email from cdan...@barnard.edu. Learn why this is important

--
You received this message because you are subscribed to the Google Groups "Archivesspace_Users_Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Archivesspace_User...@lyrasislists.org.
To view this discussion visit https://groups.google.com/a/lyrasislists.org/d/msgid/Archivesspace_Users_Group/f180cb13-c2de-4c0d-920c-4699cd77cb49n%40lyrasislists.org.

[Cristy Danford]

Hi Blake,

Thanks for your reply. Upon restart of ASpace, after approximately one minute, I see in the log the "Welcome to ArchivesSpace! You can now point your browser to http://127.0.0.1:8080 " message. I also see various "Rendered" lines for html.erb files with a "Completed 200 OK" message to follow. Also within that minute, I can successfully curl 8089 and receive DB info, but attempting to curl/wget 8080 and 8081 at any point during the 20 minute lag allows a connection, but results in 404 not found. Even after a full reboot of the server, Solr is also up and running fine during that first minute. Perhaps I'm wrong to assume that all of this indicates that ASpace is started/ready to go...

I appreciate your help!

Cristy

[Jesse Martinez]

Hi Cristy,

I've seen something like this before on rails apps running on VMs. The root cause was a very slow process of generating entropy for random number generation, which is sometimes generated using environmental noise like keyboard or mouse movements. Try installing the haveged package for your VM operating system. This will majorly boost entropy collection and should speed up your app boot times.

Jesse

To view this discussion visit https://groups.google.com/a/lyrasislists.org/d/msgid/Archivesspace_Users_Group/78a3580f-4f08-4785-ad46-f7aea293c391n%40lyrasislists.org.

--

Jesse Martinez

Senior Library Applications Developer

O'Neill Library, Boston College

jesse.m...@bc.edu

617-552-2509

he/him/his

Help us answer support requests faster by using this form:

https://library.bc.edu/systems-support

[Cristy Danford]

Jesse!!

I learned something completely new and wild today thanks to you. This solved my issue. I am so grateful and happy to have this figured out. I don't even want to think about how much time I spent on this. I hope you have an excellent weekend!

Cristy

[Jesse Martinez]

I'm so happy my suggestion worked for you, Cristy. 

Have a good weekend!

Jesse