[07:57:45] o/ [10:33:31] lunch [12:11:31] o/ [14:03:29] Happy New Year! [14:12:19] o/ [14:12:25] happy new year! :) [14:17:27] I updated my work laptop to the latest MacOS over the break, I guess I get to find out what it broke today ;P [15:35:57] \o [15:39:33] Welcome back! [15:39:42] o/ [15:46:43] not having much luck to convince java (in the opensearch centos docker image) to use our root ca... [15:48:39] :S I'm not really sure how that stuff works. It must be possible though [15:48:52] dcausse if you need SRE make an opensearch 3 image let us know, should be fairly easy since we are mirroring the opensearch 3 deb repo now [15:49:31] inflatador: thanks, it's still all kind of a manual setup at the moment [15:50:40] seems like the jvm is installed manually in /home/opensearch/jdk, could perhaps explain why it does not want to use the system wide /etc/pki/java/cacerts [15:51:23] i had mixed feelings on using the embedded jvm, but that was the default so left it. I understand why they do it, but i feel like it makes things more magical [15:52:16] I'm guessing OpenSearch 2.x is too old for y'all's experiments? We do have an image of 2.19.4 [15:53:33] trying again with forcing the trustStore... [15:53:45] inflatador: yes, we run 3.x [15:56:07] It's also a RHEL-based distro vs Debian, so if you're mounting stuff into the container the CA paths will be different [15:58:01] yes, had to use centos tools but the jvm is not installed via the system package [15:58:30] seems to work by setting the path to the trustStore, getting a 404 now [17:47:10] meh I broke relforge somehow REFERENCE: java.lang.ExceptionInInitializerError: Exception java.lang.UnsatisfiedLinkError: /usr/share/opensearch/plugins/opensearch-knn/lib/libopensearchknn_faiss_avx512.so: ... [17:56:38] :S [18:00:45] /usr/share/opensearch/plugins/opensearch-knn/lib/libopensearchknn_faiss_avx512.so: /usr/share/opensearch/data/ml_cache/pytorch/2.5.1-cpu-precxx11-linux-x86_64/libstdc++.so.6: version `GLIBCXX_3.4.20' not found... [18:02:20] hmm, i mean i know generally what that means but not how it happens. Sounds like a download for a different linux version [18:03:00] it's in "ml_cache"... tempted to wipe this out and see [18:03:09] yea that seems reasonable. Might have to reload the models? [18:03:21] better to learn this stuff now than in prod :) [18:03:27] yes... [18:19:54] AFK an hour or two, taking care of a personal mishap [18:21:28] actually this problem is not new, according to logs I see it since 2025-12-09... [18:21:43] dinner [19:07:24] lunch, back in ~40 [19:39:41] bback [20:01:20] appointment, back in ~90m [21:21:13] back [21:25:00] err, back for awhile