| Time |
Nick |
Message |
| 00:14 |
|
eady joined #evergreen |
| 02:11 |
|
jamesrf joined #evergreen |
| 05:00 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 05:47 |
|
dbwells_ joined #evergreen |
| 06:35 |
|
jamesrf joined #evergreen |
| 07:11 |
|
rjackson_isl joined #evergreen |
| 07:51 |
csharp |
jeff++ #awesome news |
| 08:17 |
|
sandbergja joined #evergreen |
| 08:20 |
|
littlet joined #evergreen |
| 08:28 |
|
bos20k joined #evergreen |
| 08:38 |
|
Dyrcona joined #evergreen |
| 08:42 |
|
mmorgan joined #evergreen |
| 08:57 |
|
aabbee joined #evergreen |
| 09:14 |
kip |
Only took a little more than a week, ouch. Still need to sticker up my carry-on. |
| 09:27 |
|
yboston joined #evergreen |
| 09:42 |
aabbee |
perhaps of interest to some: "A short list of common mistakes" in PostgreSQL https://wiki.postgresql.org/wiki/Don%27t_Do_This |
| 09:54 |
Dyrcona |
We use table inheritance, and I don't think it is a bad thing when you understand how it works. |
| 09:54 |
Dyrcona |
We probably also use between with dates and timestamps somewhere. |
| 10:04 |
csharp |
JBoyer's plan about payment types addresses one of the downsides of our use of table inheritance |
| 10:11 |
mmorgan |
Closed dates question - if a closing is set at system level, does it flow down to the branches? Or do closings need to be set explicitly at the branch level? |
| 10:12 |
* mmorgan |
believes the latter is true, but is trying to confirm. |
| 10:18 |
|
Stareagle joined #evergreen |
| 10:20 |
|
Christineb_ joined #evergreen |
| 10:22 |
|
smithy joined #evergreen |
| 10:22 |
Stareagle |
Hi Guys your help yesterday was great, thanks. Today I have another problem. I installed Evergreen using berick's script and played with the staff client. Great, but I shutdown the server before shutting down opensrf etc.......Cont |
| 10:23 |
Stareagle |
I turned on the server and used instructions I found online to startup opensrf and evergreen. Although the web page comes up I cant logon to the web staff client. What horrible thing did I do wrong? |
| 10:24 |
aabbee |
what does the javascript console say? /openils/var/log/osrfsys.log? first guess is apache or websocketd isn't running or needs to be restarted. |
| 10:27 |
miker |
mmorgan: the latter, you're correct. IIRC, there's a "also push this to all the libraries under this org unit" option |
| 10:28 |
Guest67678 |
hey hey - i'm here wokring with Stuart, we don't seem to have osrfsys.log at that path |
| 10:29 |
Dyrcona |
Stareagle: Are you using websocketd or apache-websocket? It makes a difference in startup. |
| 10:30 |
Dyrcona |
Stareagle: Also, apache requires a restart after osrf services are started or restarted. |
| 10:30 |
miker |
Dyrcona: we may be using BETWEEN on dates somewhere, but we probably shouldn't (because almost all of our date-ish things are timestamps, which are float-y, and boundaries are...difficult. interestingly, there was a recent thread on this on a PG list.) |
| 10:30 |
mmorgan |
miker: Thanks, yes, I've tested the "push to all libraries..." I believe it adds a row for each branch. |
| 10:31 |
mmorgan |
miker++ |
| 10:31 |
Guest67678 |
we restarted apache, that's when the issue of not being able to login started |
| 10:31 |
Dyrcona |
miker: I was replying to the things on the list that aabbee shared earlier, but yes, that's what the list sasy about between. |
| 10:31 |
miker |
mmorgan: it does. and we do it that way so that fines and hold processing are easier (and so you can have exceptions by removing some) |
| 10:32 |
Dyrcona |
Guest67678: Are osrf services running? What does "osrf_control -l --diagnostic" output? |
| 10:32 |
miker |
Dyrcona: ah, well, nevermind me :) ... I missed the context |
| 10:33 |
miker |
we also use rules, fwiw :) |
| 10:33 |
miker |
WORKSFORME |
| 10:33 |
Dyrcona |
:) |
| 10:33 |
Dyrcona |
Why put the features in there if people shouldn't use them, right? :) |
| 10:34 |
csharp |
@dunno add WORKSFORME WONTFIX |
| 10:34 |
pinesol |
csharp: The operation succeeded. Dunno #61 added. |
| 10:35 |
miker |
ah, bugzilla, the most considerate of all bug trackers |
| 10:36 |
csharp |
HEH |
| 10:36 |
csharp |
er.. heh |
| 10:36 |
miker |
csharp: I figured you were just channelling bugzilla there :) |
| 10:36 |
csharp |
STILLWONTFIXDAMMIT |
| 10:38 |
Guest67678 |
opensrf diagnostic output...any specific line we are looking for? |
| 10:39 |
csharp |
@dunno add "zarro boogs found" |
| 10:39 |
pinesol |
csharp: The operation succeeded. Dunno #62 added. |
| 10:40 |
csharp |
Guest67678: all lines should basically look something like this: |
| 10:40 |
csharp |
* open-ils.acq [13986] uptime=22-09:37:09 cputime=00:00:00 #drones=1/15 6% |
| 10:40 |
Guest67678 |
pcrud and permacrud services are running is that what we should be looking at? |
| 10:40 |
Guest67678 |
right, that's what we have..permacrud is at 33% if that makes a difference |
| 10:42 |
csharp |
Guest67678: a couple of us here so far have guessed that you're not able to connect to websockets (either websocketd or apache2-websockets) - have you verified that websockets are running |
| 10:42 |
Guest67678 |
i can 'login' as admin from the opac view but not in /eg/staff |
| 10:43 |
csharp |
that should also be visible in the browser console (error in red about not being able to connect to websockets) |
| 10:45 |
Dyrcona |
Guest67678: If you can login in the OPAC but not the web staff client, then it is most likely websockets. |
| 10:46 |
Guest67678 |
ah, yep that's the error |
| 10:46 |
Dyrcona |
Guest67678: You also want all of the configured services to be running. If something isn't running, you'll have problems. The open-ils.auth and open-ils.auth-internal services are more important for logging in than pcrud, and nothing actually uses permacrud, IIRC. |
| 10:47 |
csharp |
if by "berick's script" you mean the ansible installer, it may be as simple as "systemctl start websocketd-osrf" (as root)) |
| 10:49 |
Guest67678 |
ah-ha...that did it! thanks so much |
| 10:51 |
Stareagle |
Thank you one and all for your help, we really appreciate the non condescending assistance you give. Hopefully we can return the favor someday. |
| 10:54 |
Dyrcona |
berick: Has that systemd service unit from the Ansible installer been added to OpenSRF or is there a Lp bug to that effect? |
| 10:56 |
Dyrcona |
Now that I ask, I seem to recall thinking that I wouldn't use it because I don't use the nginx proxy, and I would need to modify it for my SSL cert. |
| 10:57 |
|
yboston joined #evergreen |
| 11:06 |
Bmagic |
when the action_trigger_runner.pl software runs, is it possible that it's process will complete wihle the triggers that it fired still continue on another process? Doing the "found user data" -> "collecting" -> "collected" -> etc. etc. (without action_trigger_runner.pl) ? |
| 11:17 |
|
gsams joined #evergreen |
| 11:22 |
mmorgan |
Bmagic: In my experience the process is still running until all the action trigger events cycle through from "pending" to "complete" (or whatever their final state is) |
| 11:55 |
Dyrcona |
Bmagic: I have a script for refiring collected, but not finished events. |
| 11:56 |
Dyrcona |
Check out the examples from my pre-conference presentation. |
| 12:03 |
csharp |
Bmagic: I've seen the script die off but the drones continue to work like the zombies they are |
| 12:09 |
* Dyrcona |
has seen progress continue for a few minutes after the script ends, but not for very long. |
| 12:09 |
Dyrcona |
If there are still events in a collected state a few hours later, you'll to fire them again. |
| 12:10 |
Dyrcona |
<insert verb of your choice> :) |
| 12:23 |
Dyrcona |
503 Service Unavailable -- No server is available to handle this request. Um, so how did I get the error message, then? :) |
| 12:38 |
miker |
Dyrcona: recent opensrf? you exhausted the backlog queue. |
| 12:38 |
Dyrcona |
miker: The 503? No, that came from a w3c.org server, IIRC. |
| 12:39 |
Dyrcona |
I was looking for a specification, followed a link, and that came up. |
| 12:39 |
miker |
Dyrcona: ah, continue to nevermind me, then ;) |
| 12:39 |
miker |
but recent opensrf can say that to you, too |
| 12:40 |
Dyrcona |
Good to know if I ever see it there. |
| 12:40 |
Dyrcona |
I usually see a 500 Internal Server Error. :) |
| 12:40 |
Dyrcona |
And, it's almost always my fault for forgetting to start services. |
| 12:42 |
* bshum |
joins late, but to answer the question earlier about berick's ansible script... |
| 12:42 |
StomproJ |
Does anyone remember why the check-in modifier "Clear holds shelf" is a thing? It seems like there is no harm in leaving it on all the time. Is it a possible performance issue? |
| 12:42 |
bshum |
When rebooting the server, you have to restart websocketd using the systemd call |
| 12:43 |
bshum |
Otherwise, they'll run into all the problems noted in the log :) |
| 12:43 |
bshum |
Something like "sudo systemctl restart websocketd-osrf" |
| 12:44 |
bshum |
Along with all the other restart actions |
| 12:44 |
Dyrcona |
bshum: I think that could be rectified by making sure that the order between nginx and websocketd are correct. |
| 12:45 |
Dyrcona |
With standalone websocketd, it makes no difference when it starts relative to OpenSRf services or Apache 2. |
| 12:45 |
bshum |
Right |
| 12:47 |
Dyrcona |
Systemd has a configuration option for ordering things relative to each other. It's not all terrible..... :) |
| 12:48 |
bshum |
That's not a ringing ++ but I guess they're not all winners :) |
| 12:49 |
Dyrcona |
Well, I'd prefer to not have systemd take over....everything. If it was just a replacement for init and a competitor to upstart, then I'd be much more OK with it. |
| 12:50 |
Dyrcona |
But, the intention is to make it literally the systemd and take over just about everything outside the kernel. |
| 12:51 |
Dyrcona |
I suppose if it restricted itself to emulating launchd from Mac OS X, and basically replaced init and (x)inetd, then it would be OK by me. As it is, it tries to do way too much, and I'm kind of stuck using it for now. |
| 12:53 |
* Dyrcona |
shakes his fist at the bugs in systemd-resolved for instance. |
| 12:53 |
bshum |
So I think we're leaning closer towards systemd -- now :) |
| 12:53 |
Dyrcona |
@karma systemd |
| 12:53 |
pinesol |
Dyrcona: systemd has neutral karma. |
| 12:53 |
Dyrcona |
systemd-- |
| 12:53 |
Dyrcona |
:) |
| 12:53 |
bshum |
There we go |
| 12:54 |
Dyrcona |
@karma |
| 12:54 |
pinesol |
Dyrcona: Highest karma: "gmcharlt" (4), "parts" (3), "Dyrcona" (3), "jeff" (3), and "csharp" (2). Lowest karma: "printing" (-1), "post-import-cleanup" (-1), "scrollback" (-1), "systemd" (-1), and "rhamby" (1). You (Dyrcona) are ranked 2 out of 18. |
| 12:54 |
bshum |
My real work here today is now done |
| 12:54 |
Dyrcona |
Oh, no! I can't be tied with parts.... parts-- |
| 12:54 |
Dyrcona |
Parts are too often abused in my opinion. |
| 12:55 |
Dyrcona |
And too easily so. |
| 12:55 |
Dyrcona |
But, I don't have a better solution, except to recommend using serials (not parts) for magazines. |
| 13:03 |
jeff |
yet another vendor who offers SIP2 or Patron API as their only options for patron auth. |
| 13:03 |
jeff |
Patron API it is, then! |
| 13:04 |
Dyrcona |
"Patron API?" Is that III's thing? |
| 13:05 |
jeff |
probably most commonly spoken by them, yes. |
| 13:05 |
mmorgan |
StomproJ: The Clear Holds Shelf checkin modifier will cancel shelf expired holds in addition to checking the item in, so I would think there could be performance implications if it's left on. |
| 13:06 |
jeff |
we have workstations where that checkin modifier is left on, and there are no user-reported performance issues so far, but i haven't recently read through to see what operations are being performed and if those operations may have significant impact. |
| 13:07 |
jeffdavis |
jeff: care to name-and-shame the vendor? |
| 13:07 |
jeff |
Lynda.com |
| 13:07 |
jeff |
aka LinkedIn |
| 13:07 |
jeff |
aka Microsoft |
| 13:07 |
jeff |
(last I looked) |
| 13:08 |
jeff |
jeffdavis: Did I share our Patron API shim with you at some point? I can't remember if you asked, and I can't remember what the status of that repo currently is. |
| 13:09 |
Dyrcona |
jeff: The shim sounds interesting, and yeah, we use SIP2 with Lynda.com. |
| 13:09 |
Dyrcona |
Not sure if anyone still has a subscription or however they work it. |
| 13:10 |
jeffdavis |
jeff: we talked about it but I never got the code. It would be handy to see if it's convenient to share, but I expect to add PatronAPI support for bug 1817645 in the near future. |
| 13:10 |
pinesol |
Launchpad bug 1817645 in Evergreen "Configurable patron auth and retrieval" [Wishlist,New] https://launchpad.net/bugs/1817645 - Assigned to Jeff Davis (jdavis-sitka) |
| 13:11 |
jeffdavis |
jamesrf helpfully dug up some PatronAPI documentation: https://csdirect.iii.com/sierrahelp/Content/sril/sril_patronapi.html |
| 13:11 |
Dyrcona |
Hmm... This reminds me of something.... |
| 13:11 |
jeffdavis |
but it would be good to know what fields are expected by various vendors in the response |
| 13:15 |
* jeff |
nods |
| 13:15 |
jeff |
i can confirm which ones we return and which vendors we use it with |
| 13:16 |
jeff |
Dyrcona: so does LinkedIn have access to all of your patron's data accessible via SIP2, or do you take certain steps to prevent that? :-) |
| 13:17 |
Dyrcona |
jeff: I have no idea. |
| 13:17 |
Dyrcona |
I'm not in charge of SIP configuration at that level. :) |
| 13:19 |
Dyrcona |
I would think that they would only use it for Lynda, no? Why bother to troll our patrons? |
| 13:21 |
|
yboston joined #evergreen |
| 13:44 |
|
jamesrf joined #evergreen |
| 14:05 |
jeff |
Well, right. The likelihood is low. It's more a principle of least privilege thing. |
| 14:07 |
Dyrcona |
I didn't even know the two were affiliated until you mentioned it. I'm not gonna worry about it. |
| 14:17 |
StomproJ |
mmorgan, jeff, I looked and it looks like if the copy doesn't have a status of "On Hold Shelf" then the clear hold shelf code is skipped in any case during checkin. So having it enabled all the time has no effect for checkins of copies that don't have that status. |
| 14:19 |
StomproJ |
And checking in the holdshelf happens so rarely that if those are a little slower it wouldn't matter for us. So we are going to recommend that our staff leave it on all the time. |
| 14:20 |
jeff |
sounds good! |
| 14:21 |
StomproJ |
And I'm going to open a LP ticket about possibly just doing away with it. |
| 14:22 |
jeff |
Well, I can see a mode where you normally prompt if one is encountered and and the checkin modifier causes it to skip the prompt. |
| 14:23 |
jeff |
It might also be useful to alert on the case where you're checking an item in that *isn't* shelf expired. Right now if you're configured to auto-print hold and transit slips, it just prints a slip for the same hold. |
| 14:24 |
jeff |
But I don't have specific scenarios for that well formed, so I'm going to be quiet now. :-) |
| 14:28 |
StomproJ |
Jeff, thanks, I guess if a location used the hold shelf expiration as more of a suggestion than a rule, they may like the current setup... although if you just don't check in the hold, then it doesnt' matter. |
| 14:30 |
|
Stareagle_ joined #evergreen |
| 14:32 |
Stareagle_ |
Before I go any further. I am entering some circ mods on the dev server I just installed. Is there a way to export these and other settings for when I next install evergreen? |
| 14:33 |
jeff |
With a little familiarity with the PostgreSQL command shell psql, you should be able to do that. |
| 14:35 |
Stareagle_ |
jeff: I was thinking along those lines. Time to learn the Database. Was hoping for something more automatic, but such is life :-) |
| 14:35 |
StomproJ |
Stareagle_, you may just want to use a backup of your customized database for your next install. |
| 14:38 |
Dyrcona |
Stareagle_: You can also skip the database creation steps and reuse the same database/server, assuming you're reinstalling where you've installed before and not starting over fresh. |
| 14:41 |
Dyrcona |
If you want to dump the database to restore it later, something like this should work: pg_dump evergreen -U evergreen -Fc > evergreen-$(date +%A-%Y%m%d).dmp |
| 14:43 |
Dyrcona |
Then, you can restore it like this: pg_restore -U evergreen -h localhost -C -c -d postgres evergreen-Friday-20190503.dmp |
| 14:43 |
Dyrcona |
Assuming the dump was made today. |
| 14:48 |
Stareagle_ |
Dyrcona: Thank you appreciated |
| 15:10 |
csharp |
we definitely need to flesh out some concerto stuff - or come up with a more realistic dataset for anything more than bare-bones proof-of-concept stuff |
| 15:11 |
Dyrcona |
That's why I often test with copies of production data. |
| 15:11 |
Dyrcona |
It's how I became so familiar with pg_dump and pg_restore. :) |
| 15:12 |
csharp |
yeah, same here |
| 16:21 |
|
yboston joined #evergreen |
| 16:23 |
devted |
csharp: Re: a better concerto, checkout Tonic's Condenser and Masquerade tools. |
| 16:23 |
devted |
Condenser can shrink datasets on dev and migration servers so it’s quick to provision the server and runs faster (like Concerto but on real data). i.e., work with 5% or 15% of the data while still being statistically relevant. https://www.tonic.ai/post/condenser-a-database-subsetting-tool/ |
| 16:23 |
devted |
Masquerade is a real-time scrambling proxy that can quickly scramble patrons, perfect for demo and dev servers: https://www.tonic.ai/post/masquerade-a-postgres-proxy/ |
| 16:23 |
Stareagle_ |
I tried doing a dump of the database using pg_dump evergreen -U evergreen -Fc > evergreen-$(date +%A-%Y%m%d).dmp |
| 16:23 |
devted |
These two tools could be part of the Evergreen’s community dev/demo/concerto/test conversations and workflow |
| 16:23 |
Stareagle_ |
and got the following error pg_dump: [archiver (db)] connection to database "evergreen" failed: FATAL: Peer authentication failed for user "evergreen" |
| 16:24 |
Stareagle_ |
Sorry I am new at all this |
| 16:28 |
Dyrcona |
Stareagle_: You'll have to add -h localhost to the list of command arguments. |
| 16:28 |
Dyrcona |
I copied and pasted that mostly from a backup script. |
| 16:30 |
Dyrcona |
devted: They could be, but a) I wouldn't want to force them on everyone, because I don't want to condense my datasets and b) I'd be more convinced if your posts didn't sound so much like a hard sell. :) |
| 16:31 |
Stareagle_ |
Should have known it would be simple. Need a bigger thinking cap.... |
| 16:31 |
Stareagle_ |
Suppose I am too used to MySQL and graphical interfaces |
| 16:32 |
Dyrcona |
Stareagle_: No problem. |
| 16:32 |
Dyrcona |
Well, you could install phppgadmin or use PgAdmin III, but I'm not sure how they work for dumps. |
| 16:33 |
Dyrcona |
You definitely want the -Fc on the pg_dump, btw. That writes it out in a format that pg_restore can easily use. |
| 16:33 |
devted |
Dyrcona: Sorry, I haven't used those tools and I'm not affiliated with them, which is why I said they "could be" instead of "should be" part of a better workflow. I think their use-case could be a smaller-but-still-statisticaly-relevant concerto. And when I read about them, I was excited to try them out because they could make migrations faster; dev servers would be another use-case probably, unless you had plenty of horse-power |
| 16:33 |
devted |
that it doesn't matter |
| 16:34 |
Dyrcona |
devted: Not to worry. I was just kidding about the "hard sell" though you comments to read kind of like ad copy. I realized you were being sincere, hence the smiley. |
| 16:35 |
Dyrcona |
I guess it is about time for me to call it a day. The typos are becoming more frequent. |
| 16:36 |
Dyrcona |
Stareagle_: You might find this page to be useful: https://www.postgresql.org/docs/9.5/app-pgdump.html |
| 16:36 |
devted |
Drycona: :-) cew, they're on my todo-list to checkout, but thought I'd mention them here |
| 16:37 |
* Dyrcona |
smiles at the common typo of his name that makes him sound like a special brew of Hawaiian coffee. :) |
| 16:38 |
Dyrcona |
Stareagle_: And this one, too: https://www.postgresql.org/docs/9.5/app-pgrestore.html |
| 16:38 |
Dyrcona |
If you have a different version of Postgres, you can change the version numbers in the links. |
| 16:38 |
Dyrcona |
But, not much has changed with those two commands. |
| 16:40 |
devted |
Dyrcona: doh! sorry about that, but it's late Friday afternoon and about time to sit on the beach if you have one nearby! |
| 16:40 |
Dyrcona |
@blame Friday |
| 16:40 |
pinesol |
Dyrcona: Forget it, Jake. It's just Friday. |
| 16:41 |
devted |
*laughs*... go pinesol |
| 16:42 |
Dyrcona |
Yes, a good "choice." |
| 16:45 |
Stareagle_ |
Dyrcona: Looking forward to that scotch on the front porch and once again thanks for your help. |
| 17:00 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 17:12 |
|
mmorgan left #evergreen |
| 18:07 |
|
bshum joined #evergreen |
| 18:08 |
|
troy__ joined #evergreen |
| 18:58 |
pinesol |
[evergreen|Remington Steed] LP#1782014 Fix grid path for Title in patron items out - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=b9823e3> |
