| Time |
Nick |
Message |
| 06:19 |
|
eady joined #evergreen |
| 06:31 |
pinesol_green |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 07:01 |
|
Dyrcona joined #evergreen |
| 07:05 |
|
agoben joined #evergreen |
| 07:17 |
|
rjackson_isl joined #evergreen |
| 07:17 |
|
dwgreen joined #evergreen |
| 07:33 |
|
dwgreen joined #evergreen |
| 07:59 |
|
dwgreen joined #evergreen |
| 08:44 |
|
collum joined #evergreen |
| 09:07 |
|
lsach joined #evergreen |
| 09:18 |
|
terran joined #evergreen |
| 09:40 |
|
yboston joined #evergreen |
| 09:45 |
remingtron |
@later rsulejmani See this wiki page for some Evergreen-Ansible projects (and other kinds): https://wiki.evergreen-ils.org/doku.php?id=server_installation:semi_automated&s[]=ansible |
| 09:45 |
pinesol_green |
remingtron: Error: "" is not a valid command. |
| 09:45 |
remingtron |
@later rsulejmani See this wiki page for some Evergreen-Ansible projects (and other kinds): https://wiki.evergreen-ils.org/doku.php?id=server_installation:semi_automated |
| 09:45 |
pinesol_green |
remingtron: git diff origin/hamster |
| 09:46 |
JBoyer |
I think it has to be @later tell .... |
| 09:46 |
remingtron |
@later tell rsulejmani See this wiki page for some Evergreen-Ansible projects (and other kinds): https://wiki.evergreen-ils.org/doku.php?id=server_installation:semi_automated&s[]=ansible |
| 09:46 |
pinesol_green |
remingtron: Error: "" is not a valid command. |
| 09:46 |
remingtron |
@later tell rsulejmani See this wiki page for some Evergreen-Ansible projects (and other kinds): https://wiki.evergreen-ils.org/doku.php?id=server_installation:semi_automated |
| 09:46 |
pinesol_green |
remingtron: The operation succeeded. |
| 09:46 |
remingtron |
JBoyer++ |
| 09:46 |
JBoyer |
Oh, and it doesn't like []. Didn't think of that. |
| 09:47 |
remingtron |
The things we learn! |
| 09:47 |
JBoyer |
remingtron++ # helping |
| 10:06 |
Dyrcona |
@later tell bshum [quote random] |
| 10:06 |
pinesol_green |
Dyrcona: The operation succeeded. |
| 10:07 |
Dyrcona |
And, that's why. |
| 10:07 |
bshum |
Oooh let's see what I got :D |
| 10:08 |
bshum |
@quote get 148 |
| 10:08 |
pinesol_green |
bshum: Quote #148: "-*- csharp uses force lightning on reports server" (added by gmcharlt at 09:46 AM, March 22, 2016) |
| 10:08 |
Dyrcona |
@blame [band] for [quote random] |
| 10:08 |
pinesol_green |
Dyrcona: The Version Treadmill Bullet tests their code on the LIVE SERVERS, then blames the user. SAD! for Quote #166: "< Dyrcona> Basic programmers don't die. They just GOSUB WITHOUT RETURN." (added by csharp at 10:03 AM, May 26, 2017) |
| 10:10 |
Dyrcona |
Bleh. It takes a long time to generate fines after a few days missed when you have hourly fines. |
| 10:10 |
Dyrcona |
Maybe I should write a script to check all these things i. |
| 10:10 |
Dyrcona |
in... |
| 10:10 |
Dyrcona |
It's a test server, after all. |
| 10:17 |
|
Christineb joined #evergreen |
| 10:26 |
|
yboston joined #evergreen |
| 10:50 |
JBoyer |
mmorgan, Bmagic, re: bug 1758160 I tested deleting some users today, one of which had 25K+ circs and even that was still under 10 seconds. |
| 10:50 |
pinesol_green |
Launchpad bug 1758160 in Evergreen "Deleting patrons can exceed staff client timeouts" [Undecided,Confirmed] https://launchpad.net/bugs/1758160 |
| 10:51 |
Bmagic |
huh |
| 10:51 |
Bmagic |
I'm sure hardware is a factor as well |
| 10:51 |
Dyrcona |
I'm doing a purge of 55,812 patrons on a test database with the index added. |
| 10:51 |
Dyrcona |
I'm timing it. |
| 10:51 |
JBoyer |
It can be, but most of the delay is in aging the circs, not clearing the history |
| 10:51 |
Bmagic |
our server is running on a 486 66 SX |
| 10:52 |
JBoyer |
You gotta get that DX, tho. |
| 10:52 |
Dyrcona |
Then, I'll do the same purge on a different database on the same server without the index. |
| 10:52 |
Dyrcona |
Yeah, the FPU makes all the difference! |
| 10:52 |
Bmagic |
But we do have SSD's, couldn't upgrade the Motherboard though |
| 10:53 |
Dyrcona |
What I saw with testing circulation aging this morning was a five-fold increase in speed. |
| 10:53 |
Bmagic |
had to use the adapter to get the 40 pin IDE cable to plug into the SSDs https://www.newegg.com/Product/Product.aspx?Item=9SIA7256ME5972 |
| 10:54 |
Dyrcona |
A batch size that took 25 minutes without the index took only 5 minutes with the index. |
| 10:54 |
JBoyer |
Dyrcona, how many total history entries are there? (not just for the users being deleted) We only have a little under a million here. |
| 10:54 |
Dyrcona |
JBoyer: I'm not being that scientific. |
| 10:55 |
Dyrcona |
I'm just deciding if it seems worth it to add the index in production, and so far it looks like it is. |
| 10:55 |
JBoyer |
Makes sense. |
| 10:55 |
dbwells |
One of my earliest computing memories: Me: "Dad, what does SX stand for?" Dad: "I don't know, sucky-X." |
| 10:56 |
Dyrcona |
But, it definitely speeds up aging circulations, i.e. bug 1778588. |
| 10:56 |
pinesol_green |
Launchpad bug 1778588 in Evergreen "action.purge_circulation can be extremely slow" [Undecided,New] https://launchpad.net/bugs/1778588 |
| 10:56 |
Bmagic |
dbwells++ |
| 10:56 |
Dyrcona |
The SX was basically a regular 486 with half the connections cut. |
| 10:57 |
Dyrcona |
So, yeah, Sucky-X. :) |
| 10:57 |
Bmagic |
I'm proud of my VLB video card |
| 10:57 |
Bmagic |
ISA bus with a little kick in the back |
| 10:57 |
Dyrcona |
EGA forever, baby! |
| 10:58 |
Bmagic |
https://en.wikipedia.org/wiki/VESA_Local_Bus |
| 10:58 |
Dyrcona |
Who needs more than 256 colors? |
| 10:58 |
Dyrcona |
Anyway, the main thing I want that index for is to speed up aging old circulations. |
| 10:59 |
Dyrcona |
We had 14? million of these and we're doing 50,000 a night. |
| 10:59 |
Bmagic |
Duke Nukem plays so sweet on that VLB card yall. It's a must-have for this season |
| 11:00 |
Dyrcona |
The script usually ages a few extra because of how it is written, but it typically runs about 9 hours 15 minutes. |
| 11:00 |
Dyrcona |
That's on our kick ass hardware with the NVME drives. |
| 11:01 |
Dyrcona |
If I can drop that to 1/5 or so the run time, then I can age more and be done in less than 3 years. |
| 11:04 |
Dyrcona |
dbwells: You're planning to release on Thursday, not Wednesday? |
| 11:05 |
dbwells |
Dyrcona: yes, that is the plan, on request of sandbergja for release note purposes. |
| 11:05 |
Dyrcona |
OK. Thursday works better for me, too. |
| 11:06 |
* Dyrcona |
volunteers for something... |
| 11:09 |
* dbwells |
suddenly feels like he's told that SX anecdote in here before. |
| 11:10 |
JBoyer |
Dyrcona, I don't know if you' |
| 11:11 |
dbwells |
Retelling stories, one more step on the path to oldness. Last step: not caring about retelling old stories. Not quite there yet! |
| 11:11 |
JBoyer |
re serious or not, but I don't know that an index can help with that since it's copying then deleting. :/ |
| 11:11 |
Dyrcona |
I don't know if I don't, either. |
| 11:12 |
Dyrcona |
JBoyer: Well, there are a number of factors that could have lead to the speed up, but an 80% reduction in time is nothing to sneeze at. |
| 11:12 |
JBoyer |
Oh, I think I misunderstood. |
| 11:13 |
Dyrcona |
I ran a script to age 1,000 circs. It took 25 minutes. |
| 11:13 |
Dyrcona |
I ran it again with the index added. It took 5 minutes. |
| 11:13 |
Dyrcona |
Different 1,000 circs each time. |
| 11:13 |
JBoyer |
I thought you were saying that even with this index things would be slow and that you were looking for something else to get another speed increase specific to aging circs. |
| 11:13 |
Dyrcona |
No. I think the index has helped aging circulations some how. |
| 11:14 |
JBoyer |
Definitely, since one of the triggers sets action.usr_circ_hisotry.source_circ to null as part of the process. |
| 11:15 |
Dyrcona |
I'm testing with purging (deleting) users. |
| 11:15 |
Dyrcona |
I have about 56,000 that have been deleted but didn't have their data purged. |
| 11:15 |
JBoyer |
Does the same thing in the end. :) |
| 11:16 |
Dyrcona |
I have a number of databases to play with, and I did the purge in a testing server as part of anonymizing patron data. |
| 11:16 |
Dyrcona |
It ran from about 4:00 pm friday to 10:00 monday without the index. |
| 11:17 |
* JBoyer |
is reminded that I have an A/T index bug to enter... |
| 11:17 |
Dyrcona |
That's on a decent machine. |
| 11:17 |
Dyrcona |
A former database server. |
| 11:18 |
JBoyer |
You must have a good number of history entries then. Because if that table is empty there's essentially no difference with/without the index. (That's why I was curious earlier) |
| 11:18 |
Dyrcona |
I've never counted the history entries. |
| 11:18 |
Dyrcona |
I probably should. |
| 11:19 |
Dyrcona |
Maybe next time. :) |
| 11:30 |
Dyrcona |
Almost 5 miilion: 4,971,205 user circ history entries. |
| 11:42 |
JBoyer |
Oof, yeah, having to seq scan that every single time you age a circ would be pretty bad. |
| 11:44 |
Dyrcona |
So, I think I'll go ahead and add that index to production. |
| 11:45 |
Dyrcona |
See what time the circ aging finishes tonight. |
| 11:45 |
Dyrcona |
If it speeds up a lot, I may do more than 50,000 at a time. |
| 11:57 |
|
idjit joined #evergreen |
| 12:03 |
|
khuckins joined #evergreen |
| 12:03 |
|
jihpringle joined #evergreen |
| 12:06 |
Bmagic |
Does vandelay detect the character set of the imported file? Or does it expect marc8? |
| 12:14 |
|
rsulejmani joined #evergreen |
| 12:16 |
rsulejmani |
Hi, So whenever I try and go to edit any users login info the Home Library never shows the current library it is always on the Example Library. Even after setting up the Organization units |
| 12:24 |
Dyrcona |
rsulejmani: Did you run autogen.sh? |
| 12:24 |
rsulejmani |
No How do I do that? |
| 12:25 |
Dyrcona |
As the opensrf user, you type autogen.sh, then hit Enter. :) |
| 12:25 |
rsulejmani |
And you would type that while being connected to the cd /openils folder correct? |
| 12:26 |
Dyrcona |
I don't exactly understand your question, but first sudo -i -u opensrf; then do autogen.sh |
| 12:27 |
rsulejmani |
Ok thanks I've done it. Also if I have to run autogen.sh everytime to update something is there anyway to have autogen.sh run automatically |
| 12:27 |
Dyrcona |
You have to run autogen.sh under 2 circumstances. |
| 12:27 |
Dyrcona |
1) You change organiztional unit information. |
| 12:28 |
Dyrcona |
2) The fm_IDL.xml file changes. |
| 12:28 |
rsulejmani |
Oh Ok I understand now |
| 12:28 |
Dyrcona |
These circumstances are rare in practice. |
| 12:28 |
Dyrcona |
I run it after every upgrade, just to be certain. |
| 12:29 |
rsulejmani |
Also I need some more help, Whenever I try to save the edited user info the save and the save & clone button are grayed out. |
| 12:30 |
Dyrcona |
You're missing some required field. It should be highlighted yellow or red. |
| 12:31 |
rsulejmani |
Even after filling out the fields it still shows as grayed out |
| 12:32 |
Dyrcona |
Is there a button labeled something like "show all fields?" |
| 12:33 |
kenstir |
berick: would you expect evergreen-ansible-installer to succeed after switching master to rel_3.0 ? I tried on Ubuntu 16.04 at a bad time and it failed so I reverted my snapshot. |
| 12:33 |
kenstir |
berick++ for creating automation for the people |
| 12:33 |
rsulejmani |
Yes there is |
| 12:34 |
Dyrcona |
rsulejmani: Click it and then look for empty, yellow fields. You may have a required field hidden, and that's probably a bug. |
| 12:34 |
rsulejmani |
When I clicked on it all of the fields were white, it probably is a bug |
| 12:35 |
Dyrcona |
Well, there may be some other reason related to misconfiguration. |
| 12:35 |
rsulejmani |
I did use the ansible installer |
| 12:36 |
Dyrcona |
I honestly think you should go back and install Evergreen manually from the source a few times before using automated install scripts. |
| 12:36 |
Dyrcona |
Automation is great when you know what you're doing, but not so good if you're trying to learn. |
| 12:39 |
rsulejmani |
Well I have tried but every time I do opensrf never installs and the ejabbered server never works |
| 12:39 |
rsulejmani |
That's why I choose automation |
| 12:40 |
Dyrcona |
That's understandable, but I guess it depends on what your goals are. |
| 12:41 |
Dyrcona |
Are you planning to run Evergreen in a production environment or are you just noodling around? |
| 12:41 |
berick |
kenstir: i think it should work. just be sure you match the correct opensrf version to rel_3_0. |
| 12:42 |
berick |
which I guess would be opensrf rel_3_0 |
| 12:43 |
Dyrcona |
master works, too. |
| 12:43 |
* Dyrcona |
has OpenSRF master and Evergren rel_3_0 on most of his vms. |
| 12:43 |
* berick |
nods |
| 12:44 |
kenstir |
cool, thanks! I will try again and fix anything I find |
| 12:44 |
Dyrcona |
but, yeah, rel_3_0 is the safe bet. |
| 13:04 |
|
yboston joined #evergreen |
| 13:04 |
rsulejmani |
@dyrcona Thank you for your help |
| 13:04 |
pinesol_green |
rsulejmani: Have you tried taking it apart and putting it back together again? |
| 13:05 |
Dyrcona |
You might need to restart apche2 and clear the cache in your browser. |
| 13:05 |
Dyrcona |
The web staff client is tricky. |
| 13:05 |
rsulejmani |
I have reinstalled the ubuntu os many times i think 6 or 7 and have reinstalled opensrf but everytime its the same problem |
| 13:06 |
Dyrcona |
rsulejmani: You shouldn't have to reinstall the O/S. |
| 13:06 |
Dyrcona |
You're doing this on a laptop? |
| 13:06 |
rsulejmani |
Well I resinstalled the OS to basically create a brand new copy of the files. |
| 13:07 |
rsulejmani |
Yes Im on a laptop |
| 13:07 |
Dyrcona |
You're using ubuntu 64-bit server edition? |
| 13:07 |
rsulejmani |
No Ubuntu 64 Desktop edition |
| 13:08 |
Dyrcona |
Well, in theory, it should still work, but I only ever use server edition for Evergreeen. |
| 13:08 |
|
bdljohn joined #evergreen |
| 13:08 |
Dyrcona |
32-bit gives broken libdbd, but it can be fixed. |
| 13:09 |
Dyrcona |
You should be able to clean out most of the files by doing rm -rf /openils |
| 13:09 |
Dyrcona |
There will be some Perl files left over, but they'll get replaced if you install again. |
| 13:11 |
rsulejmani |
Ok Thanks |
| 13:37 |
Dyrcona |
Jun 26 13:34:49 testing opensrf_sip[9727]: raw_transport: LOGIN ERROR: 'Failed to load ILS implementation 'OpenILS::SIP' at Sip/MsgType.pm line 918.#012' |
| 13:38 |
* Dyrcona |
is confuzzled, 'cause it's there. |
| 13:39 |
Dyrcona |
Syntax errors, of course. |
| 13:41 |
kenstir |
berick: It fails running 'npm run build-prod'. Seems to do with replacing grunt. Advice welcome. Should I add conditionals to support both npm and grunt? |
| 13:44 |
Dyrcona |
kenstir: In the Open-ILS/web/js/ui/default/staff/ did you rm -rf npm_modules before upgrading? |
| 13:47 |
kenstir |
Dyrcona: no, but the failure is due to the build-prod file missing. In recent versions of ansible-installer `grunt build` was replaced with `npm run build-prod` but that doesn't exist in rel_3_0 |
| 13:48 |
Dyrcona |
grunt was replaced with npm run build going from 3.0 to 3.1. You should probably still do grunt all. |
| 13:48 |
Dyrcona |
The ansible-installer should have nothing to do with that, but I don't use it, so I don't know. |
| 13:50 |
Dyrcona |
If ansible is doing npm run build-prod on a 3.0 branch, that's a bug in the installer, IMNSHO. |
| 13:52 |
|
khuckins_ joined #evergreen |
| 13:53 |
kenstir |
I agree, it's a bug in the installer, and there is no separate 3.0 branch of the installer. I was considering making grunt optional and not branching the installer. |
| 13:56 |
Dyrcona |
It should use grunt if the Evergreen version is not master or 3.1. I don't know how it determines that. |
| 13:57 |
bshum |
Presently it doesn't |
| 13:57 |
bshum |
It just assumes you're building master |
| 13:58 |
bshum |
Could probably be designed to handle that more gracefully I guess |
| 13:58 |
bshum |
The ansible scripting does allow us to compare variables and decide which steps to apply |
| 13:59 |
bshum |
Adding multi-version support would be similar to the experimenting I'm doing with trying to add multi-distro specific step support |
| 14:04 |
JBoyer |
It could, in fact, use the stat (or file?) module on Webpack.whatever and then run the correct command based on the result. But it currently assumes you're going to build 3.1+, and given that it's not an official tool I don't know how many knobs it needs to grow. |
| 14:04 |
bshum |
It can grow as many as berick allows us to pullrequest merge :D |
| 14:06 |
berick |
ugh, forgot about that. it does assume a master build. or at least a branch that builds the same way as master. |
| 14:06 |
kenstir |
I have a patch, I'll submit a PR |
| 14:06 |
berick |
kenstir: thanks! |
| 14:06 |
bshum |
kenstir++ |
| 14:06 |
berick |
i should probably rename the current branch to ubuntu-16.04-master |
| 14:19 |
bshum |
Maybe when: "'rel_3_0' in eg_git_branch" |
| 14:20 |
bshum |
Or hmm, no, we might have to specify the variables first. |
| 14:20 |
* bshum |
plays with this more in his abundant spare time :D |
| 14:21 |
|
abowling joined #evergreen |
| 14:29 |
kenstir |
bshum: https://github.com/berick/evergreen-ansible-installer/pull/5 |
| 14:36 |
JBoyer |
kenstir++ |
| 14:38 |
|
NFPL joined #evergreen |
| 14:40 |
NFPL |
SIP only respond with 5 holds, overdues etc when requested for that info. Is there a way to remove this limit? |
| 14:55 |
NFPL |
NVM answered my own Q. BQ field was set to 5 smh |
| 14:55 |
|
NFPL joined #evergreen |
| 15:44 |
|
khuckins joined #evergreen |
| 16:16 |
jeffdavis |
If you've got a database column can be null and the corresponding field in the fieldmapper is a link, can the link reltype be has_a or should it be might_have ? |
| 16:16 |
Dyrcona |
might_have |
| 16:17 |
Dyrcona |
But, I'm not 100% certain it makes much difference. |
| 16:21 |
jeffdavis |
The specific case I have in mind is the datatype change to asset::copy circ_as_type for bug 1743801 which we were discussing on Friday (the fix works fine in testing here, trying to avoid headaches for external services using the fieldmapper) |
| 16:21 |
pinesol_green |
Launchpad bug 1743801 in Evergreen 3.0 "web client: item status list view display issues" [High,Confirmed] https://launchpad.net/bugs/1743801 |
| 16:23 |
Dyrcona |
Let me look at something. |
| 16:29 |
|
Christineb joined #evergreen |
| 16:30 |
Dyrcona |
So, the C code mostly treats has_a and might_have the same. The link should not be virtual. |
| 16:31 |
Dyrcona |
My guess is might_have will do a right join and has_a will do a standard join, but it's not entirely clear from my cursory re-examination of the code. |
| 16:33 |
Dyrcona |
jeffdavis: It seems to make no real difference when it comes to fleshing objects. |
| 16:37 |
Dyrcona |
jeffdavis: I'd got with might_have since that implies the target might not be there. |
| 16:42 |
Dyrcona |
Filedmapper, Flattener, and Storage also don't seem to care. |
| 16:42 |
Dyrcona |
They treate has_as and might_have the same. |
| 16:42 |
* Dyrcona |
can't type. Guess I should have signed out after clocking out of work. :) |
| 17:24 |
|
khuckins_ joined #evergreen |
| 18:31 |
pinesol_green |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 23:03 |
|
eady joined #evergreen |
| 23:17 |
|
stephengwills joined #evergreen |
| 23:22 |
|
stephengwills left #evergreen |
