| Time |
Nick |
Message |
| 06:31 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 07:15 |
|
rjackson_isl joined #evergreen |
| 07:15 |
|
rjackson_isl_ joined #evergreen |
| 07:27 |
|
jvwoolf joined #evergreen |
| 08:09 |
|
kmlussier joined #evergreen |
| 08:30 |
|
rlefaive joined #evergreen |
| 08:36 |
|
mmorgan joined #evergreen |
| 08:47 |
|
rlefaive joined #evergreen |
| 08:52 |
csharp |
seeing a WSOD issue in the web client (in which the nav bar loads but no content) - the console error is "Cannot connect to offline DB: Error: http://google.github.io/lovefield/error_lookup/src/error_lookup.html?c=201&p0=Object.pkObject&p1=%5B%22asv%22%2C%221%22%5D" |
| 08:52 |
csharp |
going to that URL shows "Constraint error: (201) Duplicate keys are not allowed, index: Object.pkObject, key: ["asv","1"]" |
| 08:52 |
csharp |
just trying to log in |
| 08:53 |
csharp |
after a timeout |
| 08:53 |
csharp |
another factor might be that I just had a session from another server open in the same browser |
| 08:54 |
csharp |
yep - dead in the water even after restarting the browser |
| 08:55 |
|
bos20k joined #evergreen |
| 08:56 |
|
mllewellyn joined #evergreen |
| 08:57 |
csharp |
in Chrome's dev tools, I went to Application -> IndexedDB and deleted the database - now it works again |
| 08:58 |
csharp |
seems like the kind of thing that should be anticipated in the code |
| 08:59 |
csharp |
http://git.evergreen-ils.org/?p=Evergreen.git;a=blob;f=Open-ILS/web/js/ui/default/staff/services/lovefield.js;h=dfebc4b5b60327690f7ac22b27baab8e28722fe1;hb=refs/heads/master#l74 is where this happens |
| 08:59 |
csharp |
it just dead-ends with the error, leaving the screen blank and no indication to the user that anything is wrong |
| 09:05 |
kmlussier |
csharp: We've seen those duplicate keys errors before, and I know berick fixed one in the past. I think this one looks a little different. |
| 09:06 |
kmlussier |
Yeah, it was here. bug 1718301 |
| 09:06 |
pinesol_green |
Launchpad bug 1718301 in Evergreen "Webstaff offline DB connection failures" [High,Fix released] https://launchpad.net/bugs/1718301 |
| 09:11 |
|
mdriscoll joined #evergreen |
| 09:14 |
csharp |
kmlussier: thanks! |
| 09:15 |
csharp |
I want to reproduce it - if it's because I connected to multiple EG servers in a single browser (seems to be so, since the failure is related to actor.survey), it may not need to be considered a bug, per se |
| 09:15 |
csharp |
i.e., the answer might just be "don't do that" |
| 09:20 |
kmlussier |
csharp: I have connections open to multiple Evergreen servers in the same browsers all the time. I'm constantly doing side-by-side testing on two different servers. I don't think that should cause the problem. |
| 09:21 |
kmlussier |
csharp: When it previously happened to me, it was more likely to happen after I rebuilt my VM. |
| 09:26 |
|
yboston joined #evergreen |
| 09:26 |
|
rfrasur joined #evergreen |
| 09:32 |
|
mllewellyn joined #evergreen |
| 09:34 |
rfrasur |
Working with hatch and receipt generation. Has anyone in here (or elsewhere that you know you know about) who has inserted an image into their receipts? If so (or if you're just that awesomely intelligent), what did you include in your template code? |
| 09:36 |
|
bos20k joined #evergreen |
| 09:41 |
kmlussier |
rfrasur: I just tried it with an <img> HTML tag, but it doesn't appear to work. :( |
| 09:41 |
rfrasur |
lol, yeah, that's what we've found so far. It'll read alt text but not the img src |
| 09:47 |
kmlussier |
rfrasur: I think it's worth filing a bug on that. I'm fairly sure we have libraries that put images in their receipts. Plus, we like to have the turtle-replacement images that display on the holds, hold transit, and booking slips. |
| 09:51 |
kmlussier |
Actually, I guess the turtle-replacement images are a different matter since those aren't typically part of the receipt template. |
| 09:53 |
rfrasur |
What are they a part of?' |
| 09:54 |
rfrasur |
We have the turtle replacement things that show on the pop-up, but don't necessarily print on the slip. |
| 09:55 |
kmlussier |
rfrasur: Yes, they're identified in the code for one of the javascript files, and, as you just pointed out, they behave differently since they don't print. |
| 09:55 |
kmlussier |
I was just confusing issues earlier. No coffee yet this morning. |
| 09:56 |
mmorgan |
@coffee kmlussier |
| 09:56 |
* pinesol_green |
brews and pours a cup of El Salvador Montecarlos Tablon Crater, and sends it sliding down the bar to kmlussier |
| 09:56 |
kmlussier |
pinesol_green: Thank you for not giving me decaf. |
| 09:56 |
pinesol_green |
kmlussier: Have you confirmed your ISBN SPIDs with your service provider? |
| 09:56 |
rfrasur |
kmlussier: anything that helps suss the thing out. |
| 09:57 |
berick |
rfrasur: the print test interface includes an image. does that one work? |
| 09:57 |
rfrasur |
let's see. |
| 09:58 |
rfrasur |
berick: that's on the way. |
| 09:59 |
berick |
rfrasur: next question, are you testing on a machine w/ a valid SSL certificate? |
| 09:59 |
rfrasur |
The image does work. (I'm the go between here, sorry) |
| 09:59 |
berick |
oh, good |
| 09:59 |
rfrasur |
lol, uh huh |
| 09:59 |
rfrasur |
oh, good to the working...not the go between. |
| 10:00 |
berick |
I only talk to go-betweens! |
| 10:01 |
rfrasur |
@coffee berick |
| 10:01 |
* pinesol_green |
brews and pours a cup of Hacienda La Esmeralda, and sends it sliding down the bar to berick |
| 10:01 |
rfrasur |
You're a good person. |
| 10:02 |
berick |
heh, ok, backing up.. so the test image works? |
| 10:03 |
rfrasur |
it does, and throwing a new image in there works, but it doesn't seem to work in the actual receipts. |
| 10:04 |
berick |
ah, ok |
| 10:04 |
berick |
which receipt are you testing? |
| 10:04 |
rfrasur |
holds shelf |
| 10:05 |
csharp |
kmlussier: ah - that makes sense - I have rebuilt my test VMs DB recently |
| 10:09 |
berick |
rfrasur: i'm setting up a test to see if I can make it work.. |
| 10:11 |
rfrasur |
Excellent |
| 10:11 |
kmlussier |
@praise berick |
| 10:11 |
* pinesol_green |
berick is kind and patient to newbies |
| 10:13 |
|
Dyrcona joined #evergreen |
| 10:22 |
berick |
rfrasur: alright, it's working here. I added this just above the <table id='shelf-list-template-table'>.. in my hold shelf template. |
| 10:22 |
berick |
<img src='https://evergreen-ils.org/wp-content/uploads/2013/09/copy-Evergreen_Logo_sm072.jpg' width='140' height='24'/> |
| 10:23 |
rfrasur |
Okay, I'll report back. |
| 10:23 |
berick |
had to use an image on a server w/ a valid ssl cert -- local images won't work on OCmy dev VM. |
| 10:24 |
berick |
though we could likely tweak hatch to make that possible if needed. |
| 10:25 |
rfrasur |
I suspect it's something that should probably tweaked in hatch for the broader community for smaller implementations, etc. |
| 10:25 |
* kmlussier |
wonders how many libraries have web sites with valid ssl certs. |
| 10:26 |
kmlussier |
I know they *should* have valid ssl certificates, but I don't know if they do in reality. |
| 10:26 |
rfrasur |
kmlussier: that's my thinking as well. |
| 10:26 |
berick |
unless chrome makes you click through an ssl warning, you have a valid cert |
| 10:27 |
berick |
i dont' imagine sites would operate well w/o valid certs |
| 10:27 |
berick |
certainly not production sites |
| 10:27 |
* rfrasur |
doesn't know. |
| 10:28 |
|
mllewellyn1 joined #evergreen |
| 10:29 |
jeff |
rfrasur: if the library web site lacks a valid certificate, then the image might need to be served from the Evergreen server, where a valid certificate is presumably in place. |
| 10:29 |
kmlussier |
berick: Yes, the Evergreen server would have the valid ssl certificate. But I think some libraries might use images off their own web sites, where they may not have people logging in. |
| 10:32 |
rfrasur |
jeff: it did present from the library web site to the test receipt though. |
| 10:33 |
berick |
turns out it may not be a problem. i just printed an http:// image without issue |
| 10:33 |
berick |
i think i confused myself.. |
| 10:34 |
* rfrasur |
laughs |
| 10:34 |
berick |
the image has to come from something hatch can resolve (i.e. a resolveable hostname/IP) -- it doesn't have to be HTTPS |
| 10:35 |
|
stracey joined #evergreen |
| 10:35 |
jeff |
berick: do you know if the image is cached locally, or needs to be retrieved from the URL on every print? |
| 10:36 |
berick |
jeff: i'm fairly positive it has to be retrieved on every print |
| 10:36 |
* jeff |
remembers one library that was downloading jquery for every receipt printed |
| 10:36 |
berick |
yeah, that would be bad |
| 10:36 |
berick |
the html receipts need to be lean |
| 10:36 |
jeff |
okay, that's probably no worse than current receipts including images. |
| 10:36 |
rfrasur |
Pretty is important. Ish. |
| 10:36 |
berick |
i suspect java is not honoring my /etc/hosts file -- only reason i can think the dev vm print experiments fail |
| 10:40 |
Bmagic |
rfrasur: You should be able to convert the image into code that can be delivered right inside the template without having to link to an external site. http://dataurl.net/#dataurlmaker |
| 10:40 |
berick |
Bmagic++ # good tip |
| 10:40 |
rfrasur |
bmagic: I'll send that along and see how it goes for it. |
| 10:41 |
Bmagic |
pretty sweet thing |
| 10:41 |
jeff |
a technique that i think works in the xul client and could be tested in hatch is to use.. nevermind, interrupted and Bmagic beat me to it. :-) |
| 10:41 |
jeff |
(data: urls -- i can make no claims or assurances as to the site Bmagic mentioned) |
| 10:42 |
Bmagic |
that website will convert the image into a huge long string of text that should* work (at least it has worked for me when I needed something like that) |
| 10:42 |
kmlussier |
berick: Is the image something that should work with or without hatch? I'm testing without hatch, and, even when I see the image the preview, I don't see it in the actual slip. |
| 10:42 |
kmlussier |
I didn't try Bmagic's method yet. |
| 10:42 |
Bmagic |
I wonder if we can embed that sort of thing for the patron profile picture in the database? |
| 10:42 |
rfrasur |
No worries. Gonna give it a try...eventually. |
| 10:42 |
berick |
kmlussier: yes, it should work with or without hatch |
| 10:42 |
rfrasur |
(that's very cool, btw) |
| 10:46 |
berick |
kmlussier: in the browser, I do have to use HTTPS |
| 10:46 |
berick |
since the page is https |
| 10:46 |
csharp |
hmm - seeing an error when applying an annotated payment: https://pastebin.com/jbD3r0sQ |
| 10:47 |
csharp |
basically, error quoting string - apparently "null" is the string it's mad about |
| 10:48 |
csharp |
this is on our 3.0.1 testing server - not seeing it in 2.12.4 or on my stock master server |
| 10:48 |
Bmagic |
can I quote you csharp? |
| 10:48 |
Bmagic |
"csharp" |
| 10:48 |
Dyrcona |
Ah, null.... I was gonna say maybe you have a ' in your string. |
| 10:48 |
berick |
{value:stuff} should not inside a fieldmapper object |
| 10:49 |
berick |
*not be |
| 10:49 |
Dyrcona |
Oh, yeah! berick++ |
| 10:49 |
Dyrcona |
I was just puzzling at that. |
| 10:49 |
csharp |
Bmagic: error quoting csharp |
| 10:50 |
Dyrcona |
C# |
| 10:50 |
* csharp |
stares blankly at Dyrcona and berick |
| 10:51 |
Dyrcona |
I assume that value => '' is from a fleshed object. |
| 10:51 |
csharp |
so the problem is {"value":"csharp annotation"}? |
| 10:51 |
kmlussier |
Actually, I'm not getting any of my print template changes to appear, so it may not be a problem with the image. hmmm |
| 10:52 |
jeff |
instead of an annotation value of csharp annotation, the system attempted to create a new Fieldmapper::money::cash_payment with annotation value of {'value' => 'csharp annotation'} |
| 10:52 |
berick |
bug 1722899 |
| 10:52 |
pinesol_green |
Launchpad bug 1722899 in Evergreen "Web staff: Patron bill payment annotation dialog/modal doesn't work" [Medium,Fix released] https://launchpad.net/bugs/1722899 |
| 10:53 |
jeff |
berick++ sounded familiar |
| 10:53 |
* jeff |
pulls on another thread |
| 10:55 |
Dyrcona |
Assuming csharp has that applied, since tmcanna signed off on it, I wonder what sendPayment does. |
| 10:55 |
csharp |
berick++ # thanks - funny, the note on our issue-tracking doc even says "I wonder if this is fixed in 3.0.2?" |
| 10:55 |
berick |
jeff: and someone keeps knitting, and knitting, and knitting, and.. |
| 10:55 |
csharp |
Dyrcona: I'll have to check |
| 10:55 |
Dyrcona |
yeah. Best to make sure before looking for other causes. :) |
| 10:56 |
csharp |
nope - I think we were just assuming I'd have 3.0.2 installed by now :-/ |
| 10:56 |
Dyrcona |
I've missed important commits in production updates, and I know it can happen in testing. :) |
| 10:56 |
csharp |
this week totally got away from me |
| 10:56 |
Dyrcona |
:) |
| 10:57 |
Dyrcona |
Weeks'll do that. |
| 10:57 |
Bmagic |
weeks++ |
| 10:58 |
Bmagic |
weeks 1, days nill |
| 10:59 |
csharp |
our test server has been patched so much I'm not sure whether all of them are there |
| 10:59 |
csharp |
time to just rebuild the app servers and try again from there |
| 11:02 |
Dyrcona |
Yeap. That's why I like VMs. Just delete and build a fresh one every so often. |
| 11:10 |
csharp |
yeah, I verified that the fix was *not* applied for some reason - done now |
| 11:10 |
csharp |
and working as expected |
| 11:10 |
Dyrcona |
fixes++ |
| 11:12 |
|
Christineb joined #evergreen |
| 11:33 |
|
mllewellyn joined #evergreen |
| 12:34 |
csharp |
in the fieldmapper, if the <retreive> permission isn't specified, does that mean it's open to everyone? |
| 12:35 |
csharp |
I'm troubleshooting an issue where only the local branch's shelving locations are showing up when viewed as "Cat1" (our highest-perm cataloging role) but all are showing fine as "SuperCat" (our consortial cataloging role) |
| 12:37 |
csharp |
SuperCat has CREATE_COPY_LOCATION only - the others do not have any *_COPY_LOCATION perms |
| 12:39 |
Dyrcona |
csharp: Those permissions apply to pcrud, IIRC, and yes if retrieve is blank it's open to anyone. |
| 12:39 |
csharp |
ok - I may be barking up the wrong tree, but it feels like a perms issue |
| 12:40 |
|
rlefaive joined #evergreen |
| 12:41 |
Dyrcona |
csharp: It smells like one, and I'm not sure pcrud is being used. |
| 12:44 |
jeff |
csharp: what interface are you dealing with? |
| 12:45 |
Dyrcona |
Yeah. That will be relevant. None of the backen copy_location.retrieve methods do any perm checks. |
| 12:46 |
Dyrcona |
csharp: What is the SuperCat user's work_ou? |
| 12:49 |
Dyrcona |
Nope. Never mind that last question. retrieve.all uses the org. unit full path, so should show everything from consortium -> system -> branch, etc. |
| 12:59 |
Dyrcona |
Whoa! Just got an internal server error trying to view record summary in the staff client on master. |
| 12:59 |
|
rlefaive joined #evergreen |
| 13:00 |
Dyrcona |
Exception: OpenSRF::DomainObject::oilsMethodException 2017-12-15T12:58:43 OpenSRF::AppRequest /usr/local/share/perl/5.22.1/OpenSRF/AppSession.pm:1149 <404> Method [open-ils.circ.copy_note.retrieve.all] not found for OpenILS::Application::Circ |
| 13:00 |
Dyrcona |
Something's not quite right here. |
| 13:01 |
* Dyrcona |
wonders if the install didn't quite go right. |
| 13:07 |
Dyrcona |
Hmm... The method is defined in my Circ.pm, and I've restarted services. |
| 13:08 |
Dyrcona |
All services are running. |
| 13:11 |
jeff |
perl -c path/to/Circ.pm ? |
| 13:14 |
Dyrcona |
Yeah, gonna try that next. |
| 13:14 |
Dyrcona |
Was about to say it's not showing up in introspect |
| 13:17 |
Dyrcona |
Hm.. First line is syntax error at line 87 near }, but I don't see it. |
| 13:17 |
Dyrcona |
It is in the implementation function for copy_note.retrieve.all though. |
| 13:27 |
Dyrcona |
heh. My fault. Error in a file used by circ.pm... :) |
| 13:29 |
Dyrcona |
Weird that it didn't totally destroy circ, though. I was able to place a hold. |
| 13:31 |
Dyrcona |
And, I pick on a title that isn't holdable at example branch 1. :) |
| 14:30 |
kmlussier |
@dessert [someone] |
| 14:30 |
* pinesol_green |
grabs some Mint Chocolate Chip Cookies for phasefx |
| 14:31 |
Dyrcona |
Mm.... Mint chocolate chip cookies... Now, I want a snack. |
| 14:33 |
* kmlussier |
still has cookies sitting in her kitchen. |
| 14:34 |
* Dyrcona |
doesn't have any cookies, but could bake some, or settle for a Kind bar. |
| 14:38 |
mmorgan |
We're fresh out of kmlussier's cookies here. They were delicious! |
| 14:39 |
|
kmlussier joined #evergreen |
| 14:39 |
kmlussier |
mmorgan: I probably could have given you more. There were way too many left over. But the kids were happy. :) |
| 14:40 |
rjackson_isl |
does it seem possible that a cataloger could get an error attempting to change the owning site on a part with a duplicate key insert into the biblio.monograph_part table? The entry there is logically deleted and there are no correspooding entries in the asset_copy_part_map |
| 14:40 |
mmorgan |
Better the kids eat them than me ;-) |
| 14:41 |
rjackson_isl |
and if that seems plausible then do I fix by undeleting and adding back the row into asset.copy_part_map that isn't there? |
| 14:41 |
kmlussier |
mmorgan: Well, if it were just the kids eating them, that would be fine. |
| 14:49 |
|
mmorgan joined #evergreen |
| 14:52 |
Dyrcona |
In lieu of cookies, I found a chocolate donut. |
| 14:52 |
kmlussier |
@dessert search donut |
| 14:52 |
pinesol_green |
kmlussier: 2 found: #10: "Krispy Kreme Donuts" and #12: "Vanilla Frosted Donuts with Sprinkles" |
| 14:52 |
kmlussier |
@dessert search doughnut |
| 14:52 |
pinesol_green |
kmlussier: 1 found: #36: "Boston Cream doughnuts" |
| 14:52 |
Dyrcona |
rjackson_isl: I haven't seen that specific problem but we have had record merge fail because of duplicate parts. |
| 14:53 |
rjackson_isl |
so would it work to add in the missing asset.copy_part_map entry and undelete the monographic part entry? |
| 14:53 |
rjackson_isl |
since I can't delete the logical delete causing the duplicate |
| 14:54 |
Dyrcona |
I really don't know, unfortunately. |
| 14:54 |
rjackson_isl |
yeah - and not sure I want to experiment live! |
| 14:55 |
Dyrcona |
We solved our problem by deleting one of the duplicate part maps or whatever. |
| 14:55 |
Dyrcona |
Well, if you have a test system where you can experiment, then no real harm done. |
| 14:55 |
rjackson_isl |
problem is I am not up to speed on how they use the cataloging interface enough to pull off the test :( |
| 14:57 |
mmorgan |
rjackson_isl: I would say undeleting the monograph part and adding back the missing asset.copy_part_map should fix the immediate problem. |
| 14:57 |
rjackson_isl |
++mmorgan thanks for the input was hoping that makes sense! |
| 14:58 |
* mmorgan |
has performed minor surgery on copy_part_maps in the past. |
| 14:58 |
rjackson_isl |
yeah - I first tries a delete on monograhpic part but it was logicaly deleted already and no corresponding asset.copy_part_map entry. |
| 15:00 |
rjackson_isl |
bad thing is there are two in this reported instance and no telling how many actual! |
| 15:03 |
pastebot |
"Bmagic" at 64.57.241.14 pasted "Is this the result of looking at payment history?" (20 lines) at http://paste.evergreen-ils.org/956 |
| 15:13 |
remingtron |
Bmagic: are you running 3.0? we're upgrading next week and looking for any big road bumps to expect. |
| 15:13 |
Bmagic |
3.0.2 |
| 15:14 |
Bmagic |
JBoyer is running the same version and reported that he is not seeing that query running forever so it might just be me. I am seeing this query rarely but when it comes in, it ties up a CPU forever and I have to kill it |
| 15:14 |
Bmagic |
I had 4 of them running just now when I checked |
| 15:16 |
Bmagic |
if I could connect it to the activity on the UI that causes it, I could investigate further |
| 15:17 |
Bmagic |
I am sure that the query is generated in code with JSON or something similar making it difficult to track down the code |
| 15:18 |
berick |
Bmagic: mind re-pasting the query? |
| 15:18 |
Bmagic |
berick: I clicked the link and it's still there? Sure I'll make another one |
| 15:18 |
Bmagic |
http://paste.evergreen-ils.org/956 doesn't work for you? |
| 15:19 |
mmorgan |
seems like there are way more left joins to the same table than there should be. |
| 15:19 |
remingtron |
Bmagic: that link works for me |
| 15:20 |
remingtron |
besides that mystery, were there any surprises during the DB upgrade? |
| 15:21 |
pastebot |
"Bmagic" at 64.57.241.14 pasted "Another Example" (19 lines) at http://paste.evergreen-ils.org/957 |
| 15:21 |
berick |
Bmagic: sorry, i meant, re-paste the link |
| 15:21 |
mdriscoll |
I'm having trouble with upgrade script 3.0.1-3.0.2-upgrade-db.sql. Specifically it is failing at upgrade script 1081 with "function evergreen.container_copy_bucket_item_target_copy_inh_fkey() does not exist". I don't see where that function gets added to evergreen. There is also another missing function in the script - evergreen.asset_copy_note_owning_copy_inh_fkey(). I'm upgrading from 2.12.4 to 3.0.2. |
| 15:21 |
* berick |
nows sees the original up there |
| 15:21 |
Bmagic |
remingtron: no! The DB upgrade was perfect (after I changed a couple of lines on two of the scripts) |
| 15:23 |
jeff |
er, which lines on which scripts? :-) |
| 15:23 |
Bmagic |
remingtron: I practiced the upgrade scripts on a copy of the production DB about 5 times during the weeks leading up, so I caught 1 or 2 things |
| 15:23 |
Bmagic |
jeff: lets see |
| 15:25 |
* mdriscoll |
wonders which lines Bmagic changed in the upgrade scripts. |
| 15:25 |
Bmagic |
looking, one sec |
| 15:27 |
Bmagic |
I had to deal with 2.11.3-2.12.0 where perm codes were being added "ITEM_DEPOSIT_REQUIRED.override" and "ITEM_DEPOSIT_PAID.override" because we had already added those in our table |
| 15:28 |
Bmagic |
So I had to map the ID numbers of those two permissions over to the ID numbers that EG wants them to be |
| 15:28 |
berick |
Bmagic: my money's on /eg/staff/circ/patron/$PATRON_ID/bill_history/payments |
| 15:28 |
Bmagic |
Roughly line 403 |
| 15:28 |
Bmagic |
berick++ |
| 15:29 |
Bmagic |
berick do you think that you can see that query taking a long time on the web based staff client for any of your machines? |
| 15:29 |
Dyrcona |
mdriscoll: You're missing function is part of the base schema: ./Open-ILS/src/sql/Pg/070.schema.container.sql:CREATE OR REPLACE FUNCTION evergreen.container_copy_bucket_item_target_copy_inh_fkey() RETURNS TRIGGER AS $f$ |
| 15:29 |
Bmagic |
and this one: 2.12.6-3.0.0-upgrade-db.sql |
| 15:30 |
Bmagic |
I needed to DROP VIEW IF EXISTS action.all_circulation_combined_types; on line 26 before it was created |
| 15:30 |
Bmagic |
I had to do that because, again, we already had that view and it was slightly different than the one coming in. Different enough made postgres complain about vartypes mismatch |
| 15:31 |
Bmagic |
That was all I needed to do outside of stock upgrade scripts |
| 15:32 |
Dyrcona |
mdriscoll: So it the other function: /Open-ILS/src/sql/Pg/800.fkeys.sql:CREATE OR REPLACE FUNCTION evergreen.asset_copy_note_owning_copy_inh_fkey() RETURNS TRIGGER AS $f$ |
| 15:32 |
remingtron |
Bmagic++ #thanks for sharing! |
| 15:33 |
Dyrcona |
It's not that unusual to have to modify upgrade scripts. I almost always have to. |
| 15:35 |
berick |
Bmagic: tried a few at random and they're pretty snappy. i wonder how many rows are returned for your user ID 425135 |
| 15:35 |
Bmagic |
berick: I can confirm! Browsing to that link generates that query and the query gets stuck on postgres server |
| 15:35 |
berick |
none of my tests retunred more than a few rows |
| 15:36 |
Bmagic |
berick: well, I don't get the answer using that exact query. Which table should I look at? money.payment ? |
| 15:37 |
mdriscoll |
Dyrcona: yes, that's the other missing functions. I can add them of course and then run 3.0.1-3.0.2. |
| 15:37 |
berick |
Bmagic: meaning, it never completes? |
| 15:37 |
Bmagic |
berick: right |
| 15:37 |
Bmagic |
select * from money.billing where xact in(select id from money.grocery where usr = 425135); 0 rows |
| 15:37 |
Bmagic |
select * from money.payment where xact in(select id from money.grocery where usr = 425135); 0 rows |
| 15:39 |
berick |
Bmagic: huh. what happens if you remove the payment_ts date check from the WHERE clause? |
| 15:39 |
berick |
so it's just looking at user id |
| 15:40 |
Bmagic |
still running |
| 15:42 |
Bmagic |
still running - I don't think that made a difference |
| 15:42 |
berick |
something in that nest of joins needs help |
| 15:42 |
Bmagic |
agreed |
| 15:43 |
berick |
to be fair, i'm testing on 2.10 db |
| 15:43 |
Bmagic |
that is a lot of joins |
| 15:43 |
Bmagic |
17 left joins |
| 15:43 |
Bmagic |
and the columns selected are only from one table, lol |
| 15:44 |
berick |
but i'm around 20ms on a not-super-fast DB |
| 15:45 |
Bmagic |
I canceled it and running explain analyze on it. I'll probably have to get back to you tomorrow or the next day when it finishes |
| 15:45 |
berick |
if you're in a hurry, you could manually trim some joins, one at a time since most of them aren't needed as you said. |
| 15:46 |
Bmagic |
this might be an issue biblio.record_entry is not connected to anything |
| 15:46 |
Bmagic |
in that query |
| 15:46 |
mmorgan |
It looks like mdriscoll's function was introduced in lp 1152753, but I don't see that the function is part of the upgrade script. |
| 15:46 |
pinesol_green |
Launchpad bug 1152753 in Evergreen "Serial Units won't go into Copy Buckets" [Medium,Fix released] https://launchpad.net/bugs/1152753 |
| 15:46 |
mmorgan |
http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=f1fe22bc80f200157fe4a2f89d957a95e570282b |
| 15:46 |
pinesol_green |
mmorgan: [evergreen|Mike Rylander] LP#1152753: make it possible for serial units to be added to copy buckets - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=f1fe22b> |
| 15:48 |
Bmagic |
berick: it finished with the analyze - https://explain.depesz.com/s/Cloc |
| 15:48 |
mmorgan |
Oops. take that back. |
| 15:48 |
Bmagic |
berick: also, I redact my statement about biblio.record_entry |
| 15:51 |
jeff |
hrm. i wonder why you ran into trouble with bug 1517137 -- in theory, it was designed to not break that way. |
| 15:51 |
pinesol_green |
Launchpad bug 1517137 in Evergreen "Missing Permissions omnibus bugfix package Nov 2015" [Wishlist,Fix released] https://launchpad.net/bugs/1517137 |
| 15:51 |
Bmagic |
jeff: I was thinking that too :) |
| 15:53 |
jeff |
Bmagic: do you have record of how it failed? i might be able to replicate with simply defining those two perms with non-stock IDs and assigning them to some users/groups, but I haven't tried yet. |
| 15:54 |
Bmagic |
jeff: https://pasteboard.co/GYnfczt.png |
| 15:58 |
Bmagic |
berick: do you see that row count? Daaaaang |
| 15:58 |
berick |
yeah ;) |
| 15:58 |
Bmagic |
jeff: did that help? |
| 15:59 |
Bmagic |
jeff: I do not have a record of how it failed unfortunately. I corrected that script a few months ago and saved it "to be used during upgrade" |
| 15:59 |
jeff |
yes, a little. i was wondering what about the existing state of your database tripped it up, but that's better than nothing. :-) |
| 15:59 |
* jeff |
nods |
| 16:00 |
jeff |
it's probably too late to see if some of the technique from bug 1378829 would make that more painless. |
| 16:00 |
pinesol_green |
Launchpad bug 1378829 in Evergreen 2.12 "Typo COPY_STATUS_LONGOVERDUE.override permission" [Low,Fix released] https://launchpad.net/bugs/1378829 |
| 16:02 |
|
abowling left #evergreen |
| 16:04 |
Bmagic |
jeff: specifically http://git.evergreen-ils.org/?p=working/Evergreen.git;a=blob;f=Open-ILS/src/sql/Pg/upgrade/XXXX.data.fix_long_overdue_perm.sql;h=795f328c4f0c1290a31d60d7fb2a67330b87696b;hb=faba81f836964f3cf8639c96a881212773cc64ef ? |
| 16:07 |
jeff |
right. |
| 16:08 |
jeff |
a combination of the two approaches. |
| 16:11 |
Bmagic |
berick: money.billable_xact is defined 5 times |
| 16:11 |
|
Christineb joined #evergreen |
| 16:12 |
Bmagic |
each definition of it is followed by a group of tables that refer to that definition for a total of five groups of left joined tables |
| 16:12 |
jeff |
Bmagic: but i'm still trying to craft a scenario where the omnibus breaks. |
| 16:12 |
Bmagic |
jeff: Let me see if I can get a copy of our 2.11 database somewhere and run through vanilla scripts |
| 16:13 |
jeff |
Bmagic: no rush, but if/when you do, let me know! |
| 16:18 |
Bmagic |
berick: This smells like a bug. I can see the logic in each group considering how they are aliased "title", "record", "copy", "xact" and how it's constructed logically. But it's causing postgres to glue each group together linked back to "mp" on a loop |
| 16:18 |
Bmagic |
because each group is isolated from one another |
| 16:19 |
berick |
Bmagic: so, the query is generated by open-ils.fielder.flattened_search. it may just be it's not a good use case for it. the UI may need to move to a human-built pcrud query. |
| 16:21 |
berick |
still curious about the time differences, though, whether anyone else can reproduce. |
| 16:44 |
Bmagic |
berick: bug 1738488 for tracking purposes |
| 16:44 |
pinesol_green |
Launchpad bug 1738488 in Evergreen "Web client: patron billing history results in long running query" [Undecided,New] https://launchpad.net/bugs/1738488 |
| 16:46 |
csharp |
f1fe22bc |
| 16:46 |
pinesol_green |
csharp: [evergreen|Mike Rylander] LP#1152753: make it possible for serial units to be added to copy buckets - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=f1fe22b> |
| 16:48 |
csharp |
mdriscoll: check that you have run 2.12.6-3.0.0-upgrade-db.sql |
| 16:48 |
csharp |
maybe something in that failed |
| 16:49 |
csharp |
mdriscoll: if config.upgrade_log lacks 1063, then something went wrong and probably means you're missing more than just that |
| 16:50 |
mdriscoll |
csharp: I had a few errors with 2.12.6-3.0.0 but they were from my own stupid mistakes. I think I'll try to find another test system to run through the 2.12.4-3.0.2 scripts. |
| 16:50 |
csharp |
mdriscoll: 10-4 |
| 16:51 |
csharp |
mdriscoll: you might grep deps_block 2.12.6-3.0.0-upgrade-db.sql and make sure you aren't missing others in the list of numbered scripts - it's an all-or-nothing script though, so I would expect you're missing all of them |
| 16:51 |
csharp |
(unless you tweaked the script) |
| 16:52 |
mdriscoll |
csharp: 1063 did run. I had dropped some biblio.record_entry triggers and forgot to re-add them. Don't know if that would prevent 1063 from running completely. |
| 16:55 |
mdriscoll |
csharp: I've got this system up on 3.0.2 now, but would like to run through it all again before doing our production system in January. |
| 17:01 |
kmlussier |
OK, I'm about to sign off for two weeks. Happy holidays everyone! I'll see you in the New Year! |
| 17:02 |
|
mmorgan left #evergreen |
| 18:01 |
Bmagic |
Any thoughts on having an example birth date format next to the date of birth box on the patron registration form in Webby? |
| 18:02 |
Bmagic |
I suppose it would have to take into account what format the branch has in the settings? |
| 18:07 |
* csharp |
thinks we should have an example show |
| 18:31 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
