| Time |
Nick |
Message |
| 07:42 |
|
kworstell-isl joined #evergreen |
| 07:44 |
|
BDorsey joined #evergreen |
| 07:58 |
|
rfrasur joined #evergreen |
| 08:39 |
|
jvwoolf joined #evergreen |
| 08:39 |
|
mmorgan joined #evergreen |
| 08:40 |
jeff |
Stompro: sounds likely. I'd reach out to their support, let them know what you're seeing. |
| 08:41 |
jeff |
(you may have done that already) |
| 08:43 |
Stompro |
jeff, I did send my logs and packet dumps to my Content Cafe contact and they said they think they found the problem last night around 6:30pm cst. |
| 08:46 |
Stompro |
I also think I figured out how to enforce a timeout in AddedContent.pm that keeps things from slowing down so much. Need to keep testing that though. |
| 08:59 |
|
dguarrac joined #evergreen |
| 09:33 |
|
mantis1 joined #evergreen |
| 09:53 |
|
Dyrcona joined #evergreen |
| 10:20 |
Dyrcona |
So, I'm convinced PostgreSQL search has to go. Too many OOM errors, etc. |
| 10:27 |
miker |
https://giphy.com/gifs/thats-bait-srTYyZ1BjBtGU/fullscreen -^ |
| 10:34 |
berick |
haha |
| 10:40 |
|
RMiller joined #evergreen |
| 10:48 |
RMiller |
Help? I'm trying to finish up a migration from 3.2, and everthing looks good up until the point I go to log in. srfsh just says User login failed, like it forgot my admin user account exists from the original setup |
| 10:50 |
RMiller |
I'm confident I didn't just forget the password:) Can/should/how could I reregister the egadmin account? |
| 10:56 |
RMiller |
srfsh output starts with >Received data: "x" instead of anything hexadecimal the way it looks in the troubleshooting file. |
| 10:57 |
RMiller |
...Is that bad? |
| 11:11 |
Dyrcona |
miker++ |
| 11:11 |
berick |
RMiller: are you using "egadmin" or "admin" ? " |
| 11:11 |
berick |
... |
| 11:11 |
RMiller |
egadmin |
| 11:11 |
berick |
"x" means it didn't find the username (at I think) |
| 11:11 |
berick |
*at least |
| 11:11 |
Dyrcona |
Yeah, it's usually admin but it could be anything really. |
| 11:12 |
RMiller |
Tried with admin just for kicks, same result. I'm using the same admin account I would use in the OPAC, right? |
| 11:13 |
Dyrcona |
RMiller: Can you access the database? If so, run "select usrname from actor.usr where id = 1;" |
| 11:15 |
RMiller |
I *can* access the database, in that I have the permissions. I just don't remember what I have to do at the command line to actually do it? |
| 11:16 |
RMiller |
(I am an English teacher faking it very very hard...) |
| 11:16 |
berick |
RMiller++ |
| 11:19 |
Dyrcona |
RMiller: "psql -U evergreen -h localhost" If you need a different host or need the password, you can find it in opensrf.xml. |
| 11:20 |
RMiller |
Got it! Yes, it's "egadmin". |
| 11:23 |
|
RMiller joined #evergreen |
| 11:24 |
RMiller |
select usr.passwrd from actor.usr where id = 1; returns a password hash, so the password must still *be* there, right? |
| 11:31 |
Dyrcona |
RMiller: Passwords are actually stored in a different table now and should be automatically migrated the first time that they are used to login. |
| 11:31 |
Dyrcona |
Try "select * from actor.passwd where id = 1;" |
| 11:33 |
Dyrcona |
No, sorry: "select * from actor.passwd where usr = 1;" |
| 11:33 |
Dyrcona |
Unrelated: It's fun getting your IP blacklisted from the cloud ticketing system. |
| 11:34 |
Dyrcona |
I don't know why, unless they're detecting Chrome on Linux as some kind of bot or something. |
| 11:34 |
RMiller |
They both gave me identical output |
| 11:34 |
RMiller |
Nothing changed since 2018 :) |
| 11:34 |
Dyrcona |
RMiller: Ok, that means the user has a password. |
| 11:35 |
Dyrcona |
So, recheck your previous commands for typos. |
| 11:35 |
RMiller |
So srfsh is just not talking to the database? |
| 11:36 |
|
Christineb joined #evergreen |
| 11:36 |
Dyrcona |
srfsh doesn't talk to the database. It talks to OpenSRF services which in turn talk to the database. |
| 11:36 |
Dyrcona |
What does `osrf_control -l --diagnostic` say? |
| 11:38 |
RMiller |
Lots of lines that say uptime |
| 11:41 |
RMiller |
(Sorry, class time- afk for about 40 mintues now) |
| 11:47 |
* mmorgan |
wonders if RMiller should try restarting services |
| 11:53 |
* Dyrcona |
was thinking along similar lines to mmorgan. |
| 12:14 |
|
jihpringle joined #evergreen |
| 12:20 |
|
mantis1 joined #evergreen |
| 12:22 |
RMiller |
ejabberd status shows an error: "Failed to request certificate for biblio01, proxy.private.localhost and 11 more hosts..." Relevant? |
| 12:23 |
Dyrcona |
RMiller: Maybe. Did yous stop services during the upgrade or did you leave them running? |
| 12:25 |
RMiller |
I'm guessing that after I updated the ejabberd configuration and tested it, I probably left it running for the rest of the process |
| 12:25 |
RMiller |
I stopped OpenSRF and apache2, I'm sure |
| 12:26 |
RMiller |
ejabberd gave me an almighty hassle about the mnesia database not belonging to localhost |
| 12:28 |
RMiller |
(The message about not finding libdbi that the settings-tester.pl scripts creates doesn't apply anymore, right?) |
| 12:30 |
Dyrcona |
I never use settings-tester.pl. |
| 12:31 |
Dyrcona |
Did you upgrade or move things from server to another, or both? |
| 12:32 |
RMiller |
Upgrade, yes. Still the same server |
| 12:33 |
RMiller |
Went from 3.2 to 3.10 |
| 12:33 |
RMiller |
It kept throwing up messages about libdbi |
| 12:34 |
RMiller |
And a couple about reingesting records |
| 12:35 |
RMiller |
Which I haven't attempted yet |
| 12:41 |
Dyrcona |
RMiller: What O/S and version? The libdbi checks have pretty much been obsolete sine 2007. Assuiming by "it" you mean settings-tester.pl. |
| 12:42 |
Dyrcona |
There will be a lot of record ingests going from 3.2 to 3.10. My advice would be to make a custom db upgrade script, remove all the reignest bits, and do a full ingest after it is all done. |
| 12:43 |
RMiller |
Ubuntu Jammy. I ran all the incremental upgrade scripts to get to 3.10.0 already D: |
| 12:44 |
RMiller |
The upgrade scripts were throwing up complaints about libdbi |
| 12:46 |
RMiller |
Am I doomed? D: |
| 12:49 |
Dyrcona |
What upgrade scripts? The only upgrade scripts that I'm aware of are run via psql. |
| 12:49 |
Dyrcona |
Do a fresh install of 3.10 on jammy. Load your old database and run the db upgrade scripts on the db. Then reconfigure jammy as appropriate. |
| 12:51 |
RMiller |
Yes, the ones listed here: https://docs.evergreen-ils.org/eg/docs/latest/installation/server_upgrade.html#_upgrade_the_evergreen_code |
| 12:52 |
Dyrcona |
The upgrade for Evergreen itself is basically a new install. There is no actual upgrade path for the code. |
| 12:53 |
Dyrcona |
Did you have Evergreen installed on this system before? Did you also upgrade the O/S from a previous Ubuntu release? |
| 12:54 |
Dyrcona |
If the answer to either is "Yes," did you run the prerequisite installation for OpenSRF and Evergreen for the new O/S release? I'm starting to suspect that you have obsolete/missing packages. |
| 12:54 |
RMiller |
Yes, and yes. Upgraded from Xenial to Jammy, then OpenSRF 3.2 |
| 12:55 |
RMiller |
Then followed the Evergreen server upgrading instructions linked above |
| 12:56 |
Dyrcona |
Yeah, don't do that. Follow the README instructions for installing the new OpenSRF and Evergreen. |
| 12:56 |
RMiller |
So I should treat it as starting from scratch |
| 12:56 |
Dyrcona |
For installing the code, yes. |
| 12:58 |
Dyrcona |
I haven't looked lately, but I suspect things are hinted at in the upgrade instructions but not made explicitly clear, like having to install the prerequisites again. |
| 12:58 |
Dyrcona |
The reason that I suspect that step was missed is because you say you got libdbi errors. |
| 12:59 |
Dyrcona |
When you run osrf_control -l --diagnostic to any lines look different, something like "PID file exists but no server running?" |
| 12:59 |
Dyrcona |
s/to/do/ |
| 13:01 |
Dyrcona |
I sometimes go so far as to wipe out /openils/* when doing a big upgrade. |
| 13:01 |
RMiller |
Nope, I have a bunch of lines starting with open-ils.something [#####] uptime etc etc, and two that say router |
| 13:02 |
Dyrcona |
Well, then, it sounds like you have a working system. Does the uptime look sane? |
| 13:03 |
RMiller |
uptime matches my restarts, and cputime is all 0 |
| 13:03 |
Dyrcona |
So, the problem again is that you can't login with egadmin, right? |
| 13:04 |
Dyrcona |
Have you run authogen.sh? |
| 13:04 |
RMiller |
Yes, everything looks totally happy until it's time to log in with egadmin |
| 13:04 |
RMiller |
Yes, I've run autogen.sh after restarting the services |
| 13:05 |
Dyrcona |
You're trying to use srfsh to login or the problem with the OPAC or staff client? |
| 13:05 |
* Dyrcona |
has been doing other things and has forgotten which caused the issue. |
| 13:05 |
RMiller |
srfsh |
| 13:05 |
RMiller |
The OPAC doesn't work either, though, obviously |
| 13:06 |
Dyrcona |
Do you have another user account that you can try? |
| 13:06 |
RMiller |
Actually the OPAC shows "Login failed" before I've even tried to enter any credentials |
| 13:07 |
Dyrcona |
You generally have to restart apache2 after running autogen.sh or restart osrf services. |
| 13:07 |
Dyrcona |
That would not affect sfsh, though. |
| 13:08 |
RMiller |
right |
| 13:08 |
jeffdavis |
Is memcached running? |
| 13:08 |
RMiller |
YES |
| 13:09 |
RMiller |
Sorry. Yes :) |
| 13:09 |
jeffdavis |
I like the enthusiasm! :) |
| 13:09 |
Dyrcona |
:) |
| 13:10 |
RMiller |
When I started the troubleshooting process, systemctl stop apache2.service did not actually stop all the apache2 processes- I had to kill them manually |
| 13:12 |
Dyrcona |
I'd try shutting everything down and reinstalling OpenSRF 3.2 and Evergreen 3.10 again, following the instructions from the READMEs. |
| 13:13 |
RMiller |
Okay, so I'm backing up the evergreen database, nuking everything related to OpenSRF and Evergreen, and doing a fresh install? |
| 13:14 |
RMiller |
Thanks for your help :) |
| 13:20 |
Dyrcona |
@later tell RMiller I wouldn't necessarily "nuke everything" depending on what that means to you. |
| 13:20 |
pinesol |
Dyrcona: The operation succeeded. |
| 13:21 |
* Dyrcona |
has had mixed results upgrading Ubuntu and then upgrading Evergreen. It seems to work best with a fresh O/S install, new VM. |
| 13:22 |
Dyrcona |
It always seem to require more work after an O/S upgrade. |
| 13:23 |
Dyrcona |
typos-- |
| 13:24 |
Dyrcona |
Because I've been making more than my usual share of them today, in IRC and elsewhere. |
| 13:24 |
Dyrcona |
Oh well, at least it is Friday...."so I've got that going for me." |
| 13:27 |
Dyrcona |
Oh, sure. Look up another issue see a response with a link to another discustion: "Try post #12 see if that works." Click the link: 404 Not Found |
| 15:51 |
|
jvwoolf left #evergreen |
| 17:06 |
|
mmorgan left #evergreen |
| 20:06 |
|
kworstell-isl joined #evergreen |
| 20:53 |
|
kworstell_isl joined #evergreen |
