home.social

#systemslibrarian — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #systemslibrarian, aggregated by home.social.

  1. Like, ok, I get it, bots will scrape my site. I work in a library. Please, take my info. But this level of traffic is causing timeouts on my server and I don't like that--it reduces availability for flesh-and-blood visitors. #bots #webservers #systemslibrarian

  2. Looking at the visits logs for my small server based in Connecticut that serves up Connecticut-focused content:

    Huh, that sure is an increase over the normal background roar of bot traffic.
    Dang it, Thailand! Also, dangit Philipinnes, Taiwan, and Hong Kong.
    Anyways, off to see if there's a more granular way to block these bots than blocking the whole country. (I can't even block by IP--each request has a fresh one.) #bots #webservers #systemslibrarian

  3. That's odd. 95% of this morning's visitors to our archive server based in Western CT are from Asia. I am sure this is all entirely legitimate traffic and not bots making my life more difficult.
    (I'm fine with bots scrapping our data, it's public for a reason. I'm not fine with them DDOSing the site for hours)
    #aibots #ddos #systemslibrarian

  4. happy happy. I got to do happy low-stress techy stuff all day. I made up a stylesheet for our xml/EAD finding aids, made a printer-friendly version, and set up the web server to automagically add the stylesheet to all the .xml files in a specific directory whenever a browser asks for them. This allows me to not edit the files themselves but still have them look cute. #archives #findingaids #systemslibrarian #xml #eads #geekingout

  5. CW: I mildly broke our production server, figured out what I did, told my boss, and am currently fixing it. Happy because I'm not stressed and my current work environment is wicked supportive. Boss was like "cool, you're fixing it." cw: too much detail

    We're moving from solr 8 to 9 because the latest ArchivesSpace version demands it. Yesterday, we finished that migration. Things were a tad sluggish, probably because we were running both solr 8 and 9, since our Arclight site still used solr 8.

    I tried setting up a new Arclight core on solr 9 so I could re-index Arclight, in preparation for turning off solr 8. I accidentally deleted the website's (old, in use) Arclight core.

    So, we're moving to solr 9 a tad sooner than expected since everything needs a re-index anyway.

    #solr #oops #mybad #systemslibrarian #ibrokeit #happy

  6. CW: I updated a thing on Friday, it broke so I rolled back the update. Today I fixed the thing and updated it and now it's both updated and working! cw: geekery

    for posterity: the error message when I navigated to the .css file in a browser was

    bad message 400 reason: ambiguous uri empty segment
    and the apache config problem was the difference between the staff and public lines below:

    ProxyPass /staff
    http://localhost:8082/staff
    ProxyPassReverse /staff
    http://localhost:8082/staff
    ProxyPass /public/assets
    http://localhost:8081/assets
    ProxyPassReverse /public/assets
    http://localhost:8081/assets

    Also, omg having a dedicated test server is nice. (Remember: everyone has a test server. Ideally the test server is not the same machine as the production server.)
    #nopushfridays #updates #archivesspace #systemslibrarian

  7. CW: I updated a thing on Friday, it broke so I rolled back the update. Today I fixed the thing and updated it and now it's both updated and working! cw: geekery

    ArchivesSpace 3.5 to 4.1.1 update, my webserver config added an extra / to the end of my asset URIs, so CSS wasn't loading on the public site. Staff side was fine. #nopushfridays #updates #archivesspace #systemslibrarian