| Time |
Nick |
Message |
| 00:49 |
|
abneiman joined #evergreen |
| 01:37 |
|
pinesol joined #evergreen |
| 07:26 |
|
kmlussier joined #evergreen |
| 07:26 |
kmlussier |
Good morning evergreen! |
| 07:28 |
kmlussier |
I noticed yesterday that the EG version on Wikipedia was very out of date, so I updated the info at https://www.wikidata.org/wiki/Q613956. But it's not showing up on Wikipedia as the stable release. I'm not sure what's missing, so somebody who knows how this works may want to take a look. |
| 07:29 |
* kmlussier |
also removed the Software Freedom Conservancy from the wikidata entry and made other updates to Wikipedia, removed some libraries, updated the description to remove older elements, like Craftsman. |
| 07:29 |
kmlussier |
Somebody may want to double check my work in case I overstepped. Have a good day everyone! |
| 07:30 |
kmlussier |
parts++ |
| 07:36 |
JBoyer |
kmlussier++ |
| 07:40 |
|
collum joined #evergreen |
| 07:51 |
|
kworstell-isl joined #evergreen |
| 07:53 |
|
BDorsey joined #evergreen |
| 08:34 |
|
mmorgan joined #evergreen |
| 08:48 |
|
Dyrcona joined #evergreen |
| 09:17 |
|
dguarrac joined #evergreen |
| 10:06 |
|
rfrasur joined #evergreen |
| 10:08 |
|
sleary joined #evergreen |
| 11:14 |
Dyrcona |
Hmm.. I suspect that systemd's handling of "secure" tmp directories breaks my trick to maintain ssh auth forwarding through tmux detachment and reattach. I suspect it breaks similar setup with screen. |
| 11:15 |
Dyrcona |
What I've noticed is the link that my bash environment makes on each login ends up broken. |
| 11:16 |
Dyrcona |
It appears to have started with Ubuntu 20 on the remote host. |
| 11:18 |
|
Christineb joined #evergreen |
| 11:28 |
Dyrcona |
OK. Fixed it. Not sure I can blame systemd, really, but moving from doing the change .bash_aliases to .ssh/rc and improving the check for the existence of the socket fixed my problem. |
| 11:28 |
Dyrcona |
Basically this: https://werat.dev/blog/happy-ssh-agent-forwarding/ |
| 11:31 |
* Dyrcona |
suspects my old code was clobbering the link too often. |
| 12:10 |
|
jihpringle joined #evergreen |
| 12:52 |
|
collum joined #evergreen |
| 12:54 |
* Dyrcona |
tries restoring an evergreen database dumped from Pg 10 on Pg 15. |
| 12:59 |
Dyrcona |
Working so far.... |
| 13:08 |
Dyrcona |
I should have tried make -j20 :) |
| 13:56 |
Dyrcona |
Parallel make , ie. -j N, doesn't work with Evergreen. It works with OpenSRF. (I seem to recall coming across this before.) I should Lp it and fix it. |
| 13:57 |
|
sleary joined #evergreen |
| 14:27 |
Dyrcona |
Huh. C libraries aren't connecting to the database and I ran ldconfig again. |
| 14:30 |
Dyrcona |
Ok. The logs say "successfully connected to the database" but diagnostic says the service isn't running. This is cstore as an example. |
| 14:31 |
Dyrcona |
Dude... Over 5 minutes to connect to the database and come up.... |
| 14:39 |
csharp_ |
Dyrcona: PG config allows the same connections as the old instance? |
| 14:39 |
csharp_ |
I sometimes take the config for granted when I move versions |
| 14:40 |
csharp_ |
listen_addresses, port |
| 14:40 |
csharp_ |
but you've been around the block and have probably thought of that |
| 14:40 |
Dyrcona |
Well, it's a server that I've had configured for quite a while, and other than a pg_restore, there's nothing else going on. So much be the restore. :) |
| 14:40 |
csharp_ |
pg_hba.conf? |
| 14:41 |
Dyrcona |
It's not server config. It connected, just took a while. |
| 14:41 |
csharp_ |
ah |
| 14:42 |
Dyrcona |
The restore is probably chewing up more resources than I expect. |
| 14:42 |
csharp_ |
"wellsir" *slaps top of PG server* "this here database is slow, but she'll do the job for you" |
| 14:43 |
Dyrcona |
:) |
| 14:45 |
* Dyrcona |
is starting to get annoyed at scripts and "programs" with '_' instead of '-' in the names. :) |
| 14:45 |
* Dyrcona |
checks if Evergreen works. All the back end stuff and Apache seem happy, now, but can I login? |
| 14:47 |
Dyrcona |
And, there we go! |
| 14:48 |
csharp_ |
always satisfying |
| 14:49 |
Dyrcona |
Yeah, I'm always pleased when "request opensrf.math add 2,2" returns 4, and especially when I got to that point on muscle memory. :) |
| 14:50 |
|
mrtnnbr joined #evergreen |
| 14:50 |
Dyrcona |
I'll have to point this at Pg 15 after the restore finishes, assuming it doesn't error out. |
| 14:52 |
|
jihpringle joined #evergreen |
| 14:52 |
Dyrcona |
csharp_: I have seen it take a while for the C services to start up before, particularly with this hardware and for the first time, but not sure I've seen it take 5 minutes before today. |
| 14:52 |
Dyrcona |
OPAC seems to work. Have to try the staff client. |
| 14:54 |
Dyrcona |
And, there it is. |
| 14:55 |
Dyrcona |
Looking at the stock client reminds that one of the problems with customization these days is keeping up with the changes. We'll have to forward port our changes to Angular, now. |
| 15:08 |
Dyrcona |
I hope that we can support Pg 15 without much effort, but I haven't really made time to try it, yet. This restore is the first thing I've done. |
| 15:29 |
Dyrcona |
I'm reading the Pg 15 release notes and I don't see anything that looks like it will require us to do anything since the patches for Pg 11 - 14 went in. |
| 15:29 |
jeff |
which staff-side things do you customize locally? do you have a public repo? I'm mildly curious. |
| 15:31 |
Dyrcona |
jeff: working/collab/dyrcona/cwmars_custom_3_7_3 |
| 15:32 |
Dyrcona |
I should check that out on my new work laptop..... |
| 15:34 |
Dyrcona |
Oh... Looks like most of the interesting stuff: template changes didn't make it into that branch. |
| 15:36 |
Dyrcona |
We might be in trouble with the removal of the xml_is_well_formed function. |
| 15:38 |
Dyrcona |
jeff: We've modified the splash page to move/remove/add things. We removed the pay by credit card option from the staff client because staff should not touch credit cards. |
| 15:40 |
Dyrcona |
Oh, cute..... |
| 15:41 |
Dyrcona |
I'd like to know why xml_is_well_fomed was removed and "replaced" with xml_valid, but I'm not curious enough to go search mailing list and bug archives: https://www.postgresql.org/docs/current/xml2.html#id-1.11.7.59.5 |
| 15:44 |
Dyrcona |
it looks like we can replace the former with the latter in our code and still be safe on Pg 10. |
| 15:45 |
Dyrcona |
Looks like the changes won't work on Pg 9.6, but both 9.6 and 10 are no longer supported by the PostgreSQL community. |
| 15:49 |
Dyrcona |
if xml_is_well_formed is truly gone, then how come the restore has not blown up, yet? |
| 16:08 |
* jeffdavis |
digs through PG list archives so you don't have to: https://www.postgresql.org/message-id/2717731.1644778752@sss.pgh.pa.us and https://www.postgresql.org/message-id/E1nJdiF-00075W-EG@gemulon.postgresql.org |
| 16:10 |
jeffdavis |
also https://www.postgresql.org/message-id/1ad087a6-30aa-3371-e417-2b47956c6986@gmx.net |
| 16:10 |
Dyrcona |
jeffdavis++ |
| 16:11 |
Dyrcona |
I'm amused by "suppress link-time duplicate-symbol warnings on AIX." |
| 16:12 |
Dyrcona |
I haven't touched AIX in ... 20+ years, but my very first 'Nix account was on an AIX server at Boston University...(wow!) 30 years ago. |
| 16:15 |
Dyrcona |
Now, I wonder if the restore will "work" where a pg_upgrade reportedly failed....Doesn't mean the functions with function after the restore, though. |
| 16:15 |
Dyrcona |
s/with/will/ |
| 17:03 |
|
mmorgan left #evergreen |
| 19:21 |
|
kmlussier joined #evergreen |
| 19:23 |
kmlussier |
sandbergja++ for the wikidata update. |
| 19:26 |
kmlussier |
I still don't see what was different in your update. If you tell me how it was done, I can update the publicity checklist at https://docs.google.com/document/d/1FpCvnAgL0TiHL3L-lNWbEd7FLBzVs6UV0L1kIoN70Jw/edit?usp=sharing to provide guidance to the next person. Or anyone can, really. |
| 19:30 |
kmlussier |
Never mind. I see it now. |
| 19:32 |
jeffdavis |
sandbergja++ kmlussier++ |
| 19:42 |
|
jihpringle joined #evergreen |
