[bas@server /.../kb/cron]$ /usr/local/bin/php /.../kb/cron/cron_zkb.php
2017/01/12 10:36:33 - Starting zKB Import
2017/01/12 10:36:44 - ZKBApi: https://zkillboard.com/api/allianceID/.../
2017/01/12 10:36:44 - 0 kills were posted and 400 were skipped (400 kills fetched)
2017/01/12 10:36:44 - Timestamp of last kill: 2016-12-10 20:46:13
2017/01/12 10:36:59 - ZKBApi: https://zkillboard.com/api/allianceID/.../
2017/01/12 10:36:59 - 0 kills were posted and 400 were skipped (400 kills fetched)
2017/01/12 10:36:59 - Timestamp of last kill: 2017-01-08 10:25:24
2017/01/12 10:36:59 - Time taken = 25.760802984238 seconds.
I tried:
- clear all cache
- disable all mods
- our logs show nothing: no errors in nginx.log (server and host), php-fpm.log
- apply patch to fix zkb fetcher (class.zkbfetch.php)
- run job from admin panel and from console
I don't understand, because zkb fetcher work on other killboards on my server and class.zkbfetch.php is same on all of them.
do you get the same result if you set the last kill timestamp to a the current timestamp (to the hour)?
Since the timestamps you get are quite a bit in the past, is it not possible that all fetched mails are already present at your killboard?
ZKBApi: https://zkillboard.com/api/allianceID/.../
0 kills were posted and 400 were skipped (400 kills fetched). Timestamp of last kill: 2016-12-10 20:46:13
1st test: change timestamp from 2016-12-10 20:46:13 to 2016-12-11 00:00:00
ZKBApi: https://zkillboard.com/api/allianceID/.../
64 kills were posted and 136 were skipped (200 kills fetched). Timestamp of last kill: 2016-12-11 21:34:14
Okay, let me get this straight: During the first test, the last kill timestamp did not change at all?
Could you please PM me the alliance ID you fetched from? It still seems possible for the fetcher to get stuck under certain conditions.
I try to change timestamp to 2016-12-11 01:00:00 and saw straight behavior: fetcher works fine, but timestamp move backward until 2016-12-10 20:46:13 and after fetcher stuck.
Besides I have stucked fetcher for friend alliance too, but with another timestamp.
/.../kb/cache/error.log show nothing, server logs too
Can you verify that you are using the latest version of zKBFetch (this one).
I tried your URL on my dev KB and it worked as expected: With increasing amounts of killmails posted within one hour, the amount of mails fetched also increased:
90 kills were posted and 361 were skipped (600 kills fetched). Timestamp of last kill: 2016-12-10 20:54:46
90 kills were posted and 451 were skipped (600 kills fetched). Timestamp of last kill: 2016-12-10 21:05:19
90 kills were posted and 405 were skipped (600 kills fetched). Timestamp of last kill: 2016-12-10 21:30:49
2017/01/13 16:02:10 - 90 kills were posted and 652 were skipped (800 kills fetched)
2017/01/13 16:03:41 - 90 kills were posted and 743 were skipped (1000 kills fetched)
2017/01/13 16:04:55 - 90 kills were posted and 833 were skipped (1000 kills fetched)
2017/01/13 16:04:55 - Timestamp of last kill: 2016-12-10 23:07:39
2017/01/13 16:07:32 - 90 kills were posted and 594 were skipped (800 kills fetched)
2017/01/13 16:07:32 - Timestamp of last kill: 2016-12-10 23:32:41
Moving backward will occur, if a negative kill timestmap offset is set and your killboard does not already have the mails fetched prior to the start timestamp, and there are too many kills within that timeframe to post in one cycle. You either need to let it fetch all prior killmails until it moves forward again, or set the negative kill timestamp offset to 0.
