| Time |
Nick |
Message |
| 02:43 |
|
remingtron joined #evergreen |
| 03:55 |
|
tsbere joined #evergreen |
| 04:31 |
|
dbwells_ joined #evergreen |
| 04:34 |
|
Bmagic_ joined #evergreen |
| 04:49 |
|
wsmoak joined #evergreen |
| 07:08 |
|
agoben joined #evergreen |
| 07:24 |
|
JBoyer joined #evergreen |
| 07:52 |
|
agoben joined #evergreen |
| 08:09 |
|
collum joined #evergreen |
| 08:24 |
|
Dyrcona joined #evergreen |
| 08:46 |
|
mmorgan joined #evergreen |
| 08:53 |
|
bos20k joined #evergreen |
| 08:55 |
|
kmlussier joined #evergreen |
| 09:02 |
kmlussier |
Good morning #evergreen! |
| 09:03 |
mmorgan |
Good Morning! |
| 09:14 |
rhamby |
Good morning! |
| 09:15 |
Dyrcona |
Good morning! |
| 09:16 |
|
yboston joined #evergreen |
| 09:23 |
|
jwoodard joined #evergreen |
| 09:27 |
* Dyrcona |
should have said, "Nothing to do to save his life; call his wife in..." |
| 09:28 |
csharp |
Dyrcona: yes, we use continuous WAL archiving |
| 09:30 |
gmcharlt |
something I put together last night: https://galencharlton.com/blog/2016/10/visualizing-the-global-distribution-of-evergreen-installations-from-tarballs/ |
| 09:34 |
Dyrcona |
csharp: Thanks. I was just curious about who is using it. I'm going to do some more research into the why of doing it. It seems to have posed an issue with parallel ingest. |
| 09:36 |
Dyrcona |
A propos gmcharlt's interesting post about tarball downloads, C/W MARS did our upgrade from a custom tarball that I made. It actually worked pretty well. |
| 09:37 |
dbwells |
gmcharlt: That's pretty cool. |
| 09:37 |
gmcharlt |
yeah, I'm tempted to write a brief follow-up post to both the Koha and Evergreen ones listing caveats |
| 09:37 |
dbwells |
Kinda sad compared to the Koha map, though. |
| 09:37 |
gmcharlt |
e.g., installs from Git are likely much more common for Evergreen than Koha nowadays |
| 09:38 |
Dyrcona |
MVLC always installs from git. |
| 09:41 |
jeff |
gmcharlt: nice. :-) |
| 09:42 |
kmlussier |
gmcharlt++ |
| 09:45 |
|
maryj joined #evergreen |
| 09:59 |
Dyrcona |
dbwells | gmcharlt: I think one factor in the difference of the numbers of Evergreen vs. Koha downloads is that gmcharlt was counting Koha debian package downloads. |
| 09:59 |
Dyrcona |
Packages have a lower barrier to entry than do tarballs. |
| 10:04 |
dbwells |
I agree that that's a factor. If we feel it is a major factor, then the effort to get Evergreen packaged looks like it would be time well spent. I'm not really aware of what's involved, though. |
| 10:05 |
Dyrcona |
csharp and Emerald have made debian packages, so they would know. |
| 10:06 |
Dyrcona |
Koha has been around longer, too, so it it probably more well known, too. That is likely the largest factor. |
| 10:13 |
dbs |
Koha has that nice interactive setup & configuration ui too, IIRC, of which eg_db_config is a pale shadow |
| 10:13 |
dbs |
Dyrcona: did you get your database issues figured out? |
| 10:14 |
Dyrcona |
dbs: I stopped the ingest and services for a bit so the WAL archiver could sorta catch up. |
| 10:15 |
Dyrcona |
dbs: I'm still keeping an eye on it because the location where the WAL files are being archived fills up from time to time. |
| 10:15 |
dbs |
ah |
| 10:15 |
Dyrcona |
Doing a pg_resetxlog seems a bit drastic. |
| 10:16 |
dbs |
glad to hear you figured out a workaround -- when I was responsible for our database admin, we ran WAL and yeah, ended up carving out a lot of space for those big events :) |
| 10:16 |
dbs |
Never had to do anything quite that drastic thankfully |
| 10:29 |
|
abowling1 joined #evergreen |
| 10:31 |
|
maryj joined #evergreen |
| 10:35 |
dbs |
we upgraded from postgresql 9.1 to 9.4 over the weekend, on the off chance that that might have been part of the reason for our sluggishness (e.g. network timeout checking in a book) |
| 10:35 |
kmlussier |
dbs: Have you seen any improvement? |
| 10:36 |
dbs |
Seems unlikely--apache 2.4 + associated memory leakage seems like the most likely cause--but anything to knock out potential causes is good |
| 10:37 |
dbs |
kmlussier: haven't seen any timeouts this morning. that said, it's reading week at the university and was a long weekend for canadians so... |
| 10:38 |
kmlussier |
I don't know what reading week is, but it sounds wonderful. |
| 10:38 |
dbs |
every fall and spring, there's one week of no classes where students are supposed to be reading/doing assignments (and they often go home if they're from out of town) |
| 10:38 |
dbs |
kind of a breather |
| 10:50 |
Dyrcona |
Sounds nice for staff and students. |
| 10:58 |
dbs |
Damn, there's some cstore timeouts |
| 10:59 |
dbs |
trigger events, in this case |
| 11:10 |
dbs |
I'm more worried about staff client timeouts, so will keep an eye on this |
| 11:26 |
|
abneiman joined #evergreen |
| 11:32 |
dbs |
huh: "egweb: Context Loader error: Exception: OpenSRF::EX::ERROR 2016-10-11T15:19:55 OpenILS::WWW::EGWeb /usr/local/share/perl/5.18.2/OpenILS/WWW/EGWeb.pm:172 System ERROR: open-ils.cstore connection timed out - transaction cannot continue" |
| 11:39 |
dbs |
seems to have been about when apache processes were getting killed off due to max connections |
| 11:42 |
|
jihpringle joined #evergreen |
| 11:49 |
|
bmills joined #evergreen |
| 12:26 |
|
jeffdavis joined #evergreen |
| 12:43 |
|
sandbergja joined #evergreen |
| 12:54 |
* dbs |
notes https://www.postgresql.org/docs/9.5/static/release-9-4.html#AEN129599 "Indexes upgraded via pg_upgrade will work fine but will still be in the old, larger GIN format. Use REINDEX to recreate old GIN indexes in the new format." |
| 12:58 |
|
mmorgan1 joined #evergreen |
| 12:58 |
bshum |
dbs++ # aha, interesting.... |
| 13:01 |
|
mmorgan joined #evergreen |
| 13:05 |
dbs |
bshum++ # for making me think about indexes in the first place! |
| 13:17 |
|
krvmga joined #evergreen |
| 13:30 |
kmlussier |
Search doesn't appear to be working on webby at the moment. I'm getting zero results for all my searches. |
| 13:51 |
tsbere |
kmlussier: But is it not working because something is broken or because someone forgot to ensure there were bibs to search? ;) |
| 13:51 |
kmlussier |
The bibs are there |
| 13:52 |
kmlussier |
One of my keyword searches was for 'clifford'. At the time, I was staring at a bib record with clifford in the title. |
| 14:58 |
|
DPearl joined #evergreen |
| 15:28 |
* Dyrcona |
just got a little more dangerous: I figured out how to attach to a running Perl program and call Perl_* functions from GDB. |
| 15:30 |
|
abowling1 left #evergreen |
| 15:47 |
|
kmlussier joined #evergreen |
| 15:49 |
jeff |
pretty sure hold priority isn't going to help hbrennan's goal for "We don?t want to prevent patrons from renewing items they have just because of our staff hold." |
| 15:49 |
jeff |
but I'll confirm before suggesting as such on the list thread in question |
| 15:55 |
|
hbrennan joined #evergreen |
| 15:57 |
Dyrcona |
jeff | hbrennan: That's a neat trick. ;) |
| 15:57 |
hbrennan |
Dyrcona: Hold priority? |
| 15:58 |
jeff |
speak of the... |
| 15:58 |
jeff |
< jeff> pretty sure hold priority isn't going to help hbrennan's goal for "We don?t want to prevent patrons from renewing items they have just because of our staff hold." |
| 15:58 |
Dyrcona |
hbrennan: jeff mentioned you about 6 minutes before you entered the channel. |
| 15:58 |
kmlussier |
hbrennan: Were your ears ringing? |
| 15:58 |
jeff |
< jeff> but I'll confirm before suggesting as such on the list thread in question |
| 15:58 |
hbrennan |
whoa |
| 15:58 |
jeff |
and then yeah, you joined. |
| 15:58 |
hbrennan |
The force is strong with this one |
| 15:59 |
* mmorgan |
thinks jeff is right about the renewals. Missed that part of the question :-/ |
| 16:01 |
mmorgan |
And I think tsbere is correct about using "1" instead of "-1" though it's counterintuitive, to me anyway. |
| 16:01 |
* mmorgan |
was just looking for references in documentation. |
| 16:01 |
tsbere |
mmorgan: Was counterintuitive to me too, before I dug into the code to see how it worked just to make sure |
| 16:05 |
* mmorgan |
found this in a slideshare: |
| 16:05 |
mmorgan |
" Hold priority, if set above 0 would be assigning a priority to a group. For instance, setting it to 1 while all other groups are 0 would give that group a higher priority in the holds queue, putting them at the top of the queue every time they request an item." |
| 16:05 |
tsbere |
mmorgan: And while that sounds good the sort is still a standard numeric sort, lowest value first. <_< |
| 16:05 |
mmorgan |
but couldn't find mention in other documentation. |
| 16:06 |
mmorgan |
tsbere: Right, so documentation should reflect that. |
| 16:09 |
* mmorgan |
will open a launchpad bug. |
| 16:09 |
hbrennan |
mmorgan++ |
| 17:06 |
|
mmorgan left #evergreen |
| 21:23 |
|
wsmoak joined #evergreen |
| 21:50 |
|
dcook joined #evergreen |
