| Time |
Nick |
Message |
| 05:01 |
pinesol |
News from qatests: Failed Running pgTAP tests <http://testing.evergreen-ils.org/~live/test.42.html#2019-04-10T04:57:30,421235597-0400 -0> |
| 07:08 |
|
rjackson_isl joined #evergreen |
| 07:41 |
|
stephengwills joined #evergreen |
| 08:18 |
|
bos20k joined #evergreen |
| 08:39 |
|
mmorgan joined #evergreen |
| 08:47 |
|
stephengwills joined #evergreen |
| 08:54 |
|
Dyrcona joined #evergreen |
| 09:06 |
|
collum joined #evergreen |
| 09:35 |
|
aabbee joined #evergreen |
| 09:51 |
|
yboston joined #evergreen |
| 10:01 |
|
stephengwills joined #evergreen |
| 10:18 |
|
sandbergja joined #evergreen |
| 10:57 |
|
sandbergja joined #evergreen |
| 10:57 |
|
sandbergja_ joined #evergreen |
| 10:58 |
|
sandbergja__ joined #evergreen |
| 11:00 |
|
sandbergja___ joined #evergreen |
| 11:02 |
|
sandbergja joined #evergreen |
| 11:04 |
|
sandbergja joined #evergreen |
| 11:04 |
|
sandbergja_ joined #evergreen |
| 11:04 |
|
sandbergja joined #evergreen |
| 11:05 |
|
yboston joined #evergreen |
| 11:06 |
|
sandbergja joined #evergreen |
| 11:06 |
|
Christineb joined #evergreen |
| 11:12 |
sandbergja |
Has anybody made Evergreen work on Azure? I'm hoping to set up a small demo server (hopefully using one of BMagic's excellent docker images), and learn a bit about Azure at the same time. |
| 11:12 |
Bmagic |
I think you would be the first to use that platform |
| 11:13 |
Bmagic |
We've got it working at Google Platform and AWS. Haven't looked at Azure (yet) - Looking forward to your report |
| 11:13 |
sandbergja |
I'll see what I can find out! |
| 11:14 |
Bmagic |
Each has it's own hurdles, but I suspect it will work just fine |
| 11:15 |
sandbergja |
BMagic: I've been looking at your Google Containers slides -- do you have any documentation about the AWS version that you could share? |
| 11:16 |
bshum |
Bmagic: sandbergja: Fwiw, I used Azure to setup my VMs two Hack-A-Ways ago for test servers |
| 11:16 |
bshum |
Err, bug squashing weeks |
| 11:16 |
bshum |
Not Hackaways |
| 11:17 |
bshum |
But it's basically like any other cloud hosting option |
| 11:17 |
sandbergja |
bshum++ |
| 11:17 |
bshum |
I was using it cause I had some free credits from my Visual Studio subscription |
| 11:17 |
bshum |
And also to learn their platform |
| 11:18 |
bshum |
I didn't try setting up the dedicated separate database server deploy or their other network setup options |
| 11:18 |
bshum |
I just made them a single virtual machine and used their Ubuntu image deploy |
| 11:18 |
Bmagic |
sandbergja: oh boy, not fully. I can point you to this script: https://github.com/mcoia/mobius_evergreen/blob/master/Random/getaws_retrieve_active_ips.sh |
| 11:19 |
sandbergja |
BMagic++ |
| 11:20 |
sandbergja |
This has been super super helpful |
| 11:20 |
Bmagic |
A very very specific script for using AWS's CLI to get the Public IP's of your ELB instances, and shoving them into a cssh config file. Makes it much faster for sshing into each machine. They change frequently just as an ELB would |
| 11:23 |
Bmagic |
That's all I have "documented" - but there are so many things for AWS that are special. Sorry, I don't have them documented for public use. If you find yourself using AWS, let me know and I can try to lend a hand |
| 11:26 |
bshum |
sandbergja: In case it helps, I'd recommend that you select an Azure VM that gets you at least 4 GB of RAM, concerto doesn't like running on less for a small demo sized system. More if you are trying to make a real demo of some other database :) |
| 11:26 |
sandbergja |
bshum++ |
| 11:27 |
Bmagic |
agreed - 1 CPU and 1GB memory = crashed machine upon starting opensrf. haha |
| 11:27 |
sandbergja |
Bmagic++ |
| 11:27 |
jeff |
general advice across the board: have an awareness of the cost of the resources you're using, and have an exit strategy. don't get in a situation where you're not comfortable with burning it all down. |
| 11:27 |
Bmagic |
jeff++ # so true |
| 11:27 |
bshum |
jeff++ # wisdom from personal tragedy? :) |
| 11:28 |
jeff |
bshum: Nothing on the scale of what I'd call a tragegy, but just a general awareness of minor waste due to lack of discipline. :-) |
| 11:28 |
bshum |
Hehe |
| 11:29 |
jeff |
Some vendors will give you a forgiveness on an accidental bill, but based on what I've seen they're generally only willing to do that once. |
| 11:29 |
jeff |
I've happily never had to ask for that. |
| 11:30 |
jeff |
Especially since part of your goal is to learn the platforms in question, I recommend setting up billing alerts/alarms and looking at the other tools available for reporting on costs / budgets, etc. |
| 11:30 |
Bmagic |
yep, the money pieces can get tricky to navigate |
| 11:32 |
jeff |
It's easy to end up with an instance running in an account that you forget about, or in a region other than that which you're normally looking at, etc. |
| 11:33 |
Bmagic |
use a credit card with a $100 limit and boom, you're good. lol |
| 11:35 |
|
nfBurton joined #evergreen |
| 11:36 |
jeff |
Bmagic: ...not recommended. :-) |
| 11:40 |
jeff |
Also helpful is to have an awareness of when you're billed for something. Using AWS as an example, some things you only pay for when you're using them, other things you pay for any time they're provisioned, certain other things you pay for only when you're NOT using them... :-) |
| 11:48 |
sandbergja |
jeff++ |
| 11:48 |
berick |
shout out to amazon lightsail for small vm's https://aws.amazon.com/lightsail/pricing/ |
| 11:51 |
|
khuckins joined #evergreen |
| 11:58 |
|
jihpringle joined #evergreen |
| 11:59 |
jeff |
berick++ I forget about Lightsail. There are some useful bundles there, especially if you're looking for simpler pricing. |
| 12:25 |
jeff |
surprising Lightsail API quirk: granting "Read" access gives a bit more than you'd probably expect, since it gives you access to the default private SSH key which you can then probably use to log in to an instance as an administrative user with much-more-than-Read permissions. |
| 12:27 |
jeff |
I mean, it's still only granting Read access to the API itself... but... :-) |
| 12:41 |
bos20k |
Dyrcona and phasefx: Just wanted to let you know that OpenSRF bugs 1824181 and 1824184 have been filed that allow and make use of delayed execution subroutines for logging in OpenSRF. |
| 12:41 |
pinesol |
Launchpad bug 1824181 in OpenSRF "Allow first argument to logger to be string or subroutine" [Undecided,New] https://launchpad.net/bugs/1824181 |
| 12:41 |
pinesol |
Launchpad bug 1824184 in OpenSRF "Change potentially slow log statements to subroutines" [Undecided,New] https://launchpad.net/bugs/1824184 |
| 12:41 |
bos20k |
Dyrcona++ |
| 12:41 |
bos20k |
phasefx++ |
| 12:46 |
pinesol |
[evergreen|Jason Etheridge] lp1145213 avoid race condition in bib merge test - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=bbe8a92> |
| 12:52 |
|
yboston joined #evergreen |
| 12:58 |
* mmorgan |
is wondering if a Best-Hold selection sort order can help with this situation: |
| 12:58 |
mmorgan |
Library A's book is popular at Library B and always has holds for pickup there. |
| 12:58 |
mmorgan |
A hold for pickup at Library C is placed, but is not filled because of the holds for pickup at Library B. |
| 12:59 |
mmorgan |
Is there a way to prefer Library C's hold when it remains unfilled for a period of time? |
| 12:59 |
Dyrcona |
Library B should buy a copy. |
| 12:59 |
mmorgan |
Dyrcona: Understood. But my question remains :) |
| 13:00 |
Dyrcona |
I don't know. I've never messed with hold sort order. |
| 13:01 |
jeff |
we use request order (formerly "FIFO") |
| 13:01 |
mmorgan |
I am pretty sure the answer to my question is no, but I want to make sure I haven't missed anything. |
| 13:03 |
mmorgan |
jeff: We use proximity, which works well, generally, reducing transits while filling holds in a reasonable order, but there are edge cases. |
| 13:04 |
mmorgan |
jeff: Do you use holds go home? |
| 13:06 |
jeff |
We do not. |
| 13:08 |
mmorgan |
We do use holds go home so a library's patrons will have priority for that library's items. |
| 13:14 |
|
sandbergja joined #evergreen |
| 13:15 |
|
sandbergja_ joined #evergreen |
| 13:29 |
|
remingtron joined #evergreen |
| 14:08 |
phasefx |
bos20k++ |
| 14:11 |
bos20k |
phasefx: Thanks :) My change to the logger is extremely simple but gets the job done. It is very possible that there is some other really cool way to do it that would be better but this is what I found in my research and testing. |
| 14:20 |
miker |
mmorgan: that's the tension between "reduce transits" and "fill in order" when there are few copies of a popular title (obviously). there's no way today to say "ok, push holds older than X in front of really new holds", but I /think/ that's what you want, approximately? |
| 14:21 |
jeff |
"favor prox unless hold being considered is older than X" or something similar? |
| 14:21 |
miker |
(and, of course, "practice fair collection development" :) ) |
| 14:23 |
miker |
jeff: well, cast in terms of best-hold sort order, something like: ORDER BY request_time > $setting_value DESC, prox, ... |
| 14:24 |
miker |
is what I was thinking |
| 14:24 |
* jeff |
nods |
| 14:24 |
miker |
(setting value is owned by copy circ lib, I guess?) |
| 14:24 |
jeff |
realized i was describing it in terms that assumed default config |
| 14:25 |
jeff |
(and not more generally applicable terms) |
| 14:25 |
miker |
and that ORDER BY is constructed based on best-hold-sort-order config with "old-hold" being a new sort axis |
| 14:34 |
|
nfBurton joined #evergreen |
| 14:40 |
jeffdavis |
hmm, I really killed the discussion on bug 1822875, didn't I |
| 14:40 |
pinesol |
Launchpad bug 1822875 in Evergreen "OPAC titles are truncated when query string present" [Undecided,Confirmed] https://launchpad.net/bugs/1822875 |
| 15:05 |
* mmorgan |
returns and reads up |
| 15:07 |
mmorgan |
jeff: miker: Yes, something like "favor prox unless hold being considered is older than X" would help when the hold queue needs a little help getting realigned |
| 15:11 |
mmorgan |
And practicing fair collection development is always important, but there are always edge cases that pop up. |
| 15:16 |
mmorgan |
Edge cases that pop up after flying under the radar for a while, that is. |
| 15:35 |
Dyrcona |
So, I found what looks like an unconverted XUL copy template in this user's web staff client copy templates, and this user has no staff client template by that name. |
| 15:36 |
Dyrcona |
i.e. No XUL template by that name. |
| 15:46 |
mmorgan |
Dyrcona: Just one template that looked unconverted? Other templates in the usr setting look to have been converted? |
| 15:46 |
Dyrcona |
Yes, and the unconverted one does not exist in the XUL templates for this user. |
| 15:51 |
* mmorgan |
wonders what happens if you export templates from the xul client and import them into the web client. Do they get converted when imported into the web client? |
| 15:54 |
Dyrcona |
Interesting... I see something similar but different on a test database that was updated a few weeks ago. |
| 15:54 |
Dyrcona |
None of this user's webstaff copy template were properly converted, and the JUVINILE template does not exist in the XUL templates. |
| 15:55 |
Dyrcona |
I'll have to do more digging to see if they are even this user's templates... |
| 15:59 |
Dyrcona |
Did web staff copy templates have to be converted prior to 3.0? Could be this person used the web client on 2.12. |
| 16:04 |
mmorgan |
Dyrcona: bug 1691269 enabled converting xul to web templates, released in 3.0.2 |
| 16:04 |
pinesol |
Launchpad bug 1691269 in Evergreen "web client: copy templates created on XUL not displayed" [Medium,Fix released] https://launchpad.net/bugs/1691269 |
| 16:05 |
|
yboston joined #evergreen |
| 16:30 |
|
sandbergja joined #evergreen |
| 17:01 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 17:04 |
|
mmorgan left #evergreen |
| 17:36 |
|
sandbergja joined #evergreen |
| 18:04 |
|
stephengwills joined #evergreen |
| 19:05 |
|
sandbergja joined #evergreen |
| 20:30 |
|
remingtron_ joined #evergreen |
| 20:30 |
|
JBoyer_alt joined #evergreen |
| 20:31 |
|
pinesol` joined #evergreen |
| 21:45 |
|
sandbergja joined #evergreen |
| 23:25 |
|
sandbergja joined #evergreen |
