| Time |
Nick |
Message |
| 07:08 |
|
collum joined #evergreen |
| 07:56 |
|
collum joined #evergreen |
| 08:42 |
|
mmorgan joined #evergreen |
| 09:02 |
csharp_ |
jeffdavis++ |
| 09:02 |
* csharp_ |
opens bug: Wishlist: End All Wars |
| 09:05 |
|
Dyrcona joined #evergreen |
| 09:43 |
eeevil |
gitweb seems sad... :( |
| 09:50 |
Dyrcona |
eeevil: Looks like Apache crashed. I'm rebooting the git server. |
| 09:50 |
Dyrcona |
Looks good to me now. |
| 10:04 |
csharp_ |
stupid bots |
| 10:22 |
|
mmorgan1 joined #evergreen |
| 10:27 |
|
sandbergja joined #evergreen |
| 10:48 |
|
sandbergja joined #evergreen |
| 11:30 |
Dyrcona |
@blame bots |
| 11:30 |
pinesol |
Dyrcona: bots WILL PERISH UNDER MAXIMUM DELETION! DELETE. DELETE. DELETE! |
| 11:33 |
|
Christineb joined #evergreen |
| 12:06 |
* sleary |
points a sonic screwdriver at pinesol |
| 12:14 |
Dyrcona |
eeevil: Do you think authority_authority_linker.pl would work in parallel? |
| 12:16 |
eeevil |
Dyrcona: today, pre-THAT-patch, I assume you mean? hrm... it should be able to in theory, because it doesn't actually change any headings, it just adds $0s to variant term fields ... BUT, that doesn't mean we're not doing the naive thing and just letting it do an auth reingest that triggers (useless) browse entry updates |
| 12:27 |
Dyrcona |
Well, I got it down to 22 hours on my test system by doing batches of 10,000 records at a time. Doing the --all option took almost 72 hours. |
| 12:28 |
Dyrcona |
I've been asked about running both the authority to authority linker and authority control fields in production, so I'm trying to estimate how long it will take. |
| 12:29 |
Dyrcona |
I am going to test with the parallel browse ingest branch this time. |
| 13:32 |
|
mixo joined #evergreen |
| 13:38 |
|
jonadab joined #evergreen |
| 13:40 |
mixo |
Hello! I am trying to generate reports but when I click "View Report Output" I get an ERR_TOO_MANY_REDIRECTS error. |
| 13:40 |
mixo |
I am using the nginx reverse proxy from the OpenSRF installation guide and I am using the osrf-ws-http-proxy configuration file for it. |
| 14:01 |
mixo |
Nothing error in osrfsys.log |
| 14:01 |
Bmagic |
mixo: Does your apache VirtualHost block have this line: ServerName localhost:443 |
| 14:02 |
Bmagic |
specifically inside block <VirtualHost *:7443> |
| 14:03 |
mixo |
Yes, I have <VirtualHost *:7443> |
| 14:03 |
mixo |
DocumentRoot "/openils/var/web" |
| 14:03 |
mixo |
ServerName localhost:7443 |
| 14:03 |
mixo |
ServerAlias 127.0.0.1:7443 |
| 14:03 |
mixo |
SSLEngine on |
| 14:03 |
mixo |
SSLProxyEngine on # required for ErrorDocument 404 on SSL connections |
| 14:03 |
mixo |
SSLHonorCipherOrder On |
| 14:03 |
mixo |
SSLCipherSuite ECDHE-RSA-AES256-SHA384:AES256-SHA256:RC4:HIGH:!MD5:!aNULL:!EDH:!AESGCM |
| 14:03 |
Bmagic |
I bet that's your issue: change 7443 to 443 inside the block |
| 14:04 |
mixo |
but nginx listens to 443 port |
| 14:04 |
Bmagic |
but keep <VirtualHost *:7443> with 7443. Just concerned about "ServerName localhost:7443" and "ServerAlias 127.0.0.1:7443" <- make those 443 instead |
| 14:05 |
Bmagic |
mixo: understood, it's weird, I know, but that's what your setup will likely need |
| 14:05 |
mixo |
Ok |
| 14:09 |
mixo |
I have now "The requested URL was not found on this server. ". I will create a report from scratch. Thank you. |
| 14:30 |
Dyrcona |
huh. authority_authority_linker seems to be running faster with a 3.15.3 database than it did with a 3.7.4 database. Same database server and identical vm specs for the Evergreen side of things. |
| 14:31 |
Dyrcona |
Maybe it's the parallel browse ingest patch, but didn't think that would affect the authorities. |
| 14:35 |
Dyrcona |
Is it recommended anywhere to run the authority_authority_linker or authority_control_fileds daily? Someone said it in a local email, and I don't remember seeing that recommended anywhere. And, given that ours takes 3 days to run on all records, that sounds right out. |
| 14:41 |
gmcharlt |
if you're creating or updating authority records frequently, sure (though you'd need to shoehorn in something to identify the records to set --start_id and --end_id appropriately at the moment since there's --since switch or the ilke) |
| 14:42 |
gmcharlt |
* there's _no_ --since switch |
| 14:44 |
Bmagic |
mixo: you've fixed the redirect loop, so, like you summized, you need to get a successful report output now. |
| 14:46 |
mixo |
Yes, I will try create report. Thank you again. |
| 14:50 |
Dyrcona |
gmcharlt: Thanks! authority_control_fields has a --days_back switch. Maybe we could add something like that to the authority_authority_linker? Might need more changes than it seems at first. |
| 14:54 |
gmcharlt |
yeah, that should work, or at least be a good start |
| 14:55 |
gmcharlt |
though we'd need to double-check additional selection logic is needed to also look at authorities that might link _to_ the new or edited authorities |
| 15:05 |
eeevil |
hopefully we can use the main (probably simple?) heading from the new one to find (old, unlinked) variants that need to be linked |
| 15:16 |
Dyrcona |
eeevil: The branch on Lp 2091748 could use a rebase. There's a conflict in Open-ILS/src/sql/Pg/030.schema.metabib.sql. It looks simple to resolve, but I might have done it wrong when applied to 3.15.3. |
| 15:16 |
pinesol |
Launchpad bug 2091748 in Evergreen "Concurrent browse entry update/insert" [Wishlist,New] https://launchpad.net/bugs/2091748 - Assigned to Jason Stephenson (jstephenson) |
| 15:16 |
Dyrcona |
I just cherry-picked it to main and it has the same conflict. |
| 15:17 |
eeevil |
can do |
| 15:26 |
eeevil |
Dyrcona: looks like LP#1695911 changed the ON CONFLICT UPDATE a little. did you resolve the conflict by just removing the whole conflicting block? (that would be my resolution, though I'll need to go look at the reifier to see if we need to add in the SUBSTRING() from 1695911 |
| 15:32 |
Dyrcona |
I did remove the conflicting block. |
| 15:33 |
Dyrcona |
It looked like that was the solution based on the diff. |
| 15:38 |
eeevil |
actually ... I'll add the substring to the _updates insert |
| 15:39 |
eeevil |
Dyrcona: https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/miker/lp-2091748-concurrent-browse-entry-rebase-2025-08-06 is rebased and updated |
| 15:39 |
|
redavis joined #evergreen |
| 15:40 |
eeevil |
so that both addresses the rebase conflict /and/ brings back the SUBSTRING() calls |
| 15:54 |
Dyrcona |
eeevil++ |
| 15:54 |
Dyrcona |
I'll update my test system. |
| 17:04 |
|
mmorgan1 left #evergreen |
