Grommunio Archives shows only today's mails

ozy

As the post suggests, groarchive only shows mails of current day, so it looks like indexes are lost/rebuilt on every day.

I've lurked around and noticed that the sphinx index is really too small:

email:/var/lib/grommunio-archive/sphinx # ls -alh
total 428K
drwxr-x--- 2 groarchive groarchive 4,0K mars  20 10:15 .
drwxr-x--- 8 groarchive groarchive   81 mars   6 20:32 ..
-rw------- 1 groarchive sphinx        8 mars  20 10:15 binlog.001
-rw------- 1 groarchive sphinx        0 mars  19 23:46 binlog.lock
-rw------- 1 groarchive sphinx       11 mars  20 10:15 binlog.meta
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:05 dailydelta1.spl
-rw-r--r-- 1 groarchive groarchive 1,7K mars  20 10:05 delta1.spa
-rw-r--r-- 1 groarchive groarchive  94K mars  20 10:05 delta1.spd
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:05 delta1.spe
-rw-r--r-- 1 groarchive groarchive 1,2K mars  20 10:05 delta1.sph
-rw-r--r-- 1 groarchive groarchive  71K mars  20 10:05 delta1.spi
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:05 delta1.spk
-rw------- 1 groarchive sphinx        0 mars  20 10:05 delta1.spl
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:05 delta1.spm
-rw-r--r-- 1 groarchive groarchive 173K mars  20 10:05 delta1.spp
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:05 delta1.sps
-rw-r--r-- 1 groarchive groarchive    0 mars  20 02:30 main1.spl
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 note1.spa
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.spd
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.spe
-rw-r--r-- 1 groarchive groarchive  872 mars  20 10:15 note1.sph
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.spi
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 note1.spk
-rw------- 1 groarchive sphinx        0 mars  20 10:15 note1.spl
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 note1.spm
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.spp
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.sps
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 note1.tmp.tmps
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 tag1.spa
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.spd
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.spe
-rw-r--r-- 1 groarchive groarchive  871 mars  20 10:15 tag1.sph
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.spi
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 tag1.spk
-rw------- 1 groarchive sphinx        0 mars  20 10:15 tag1.spl
-rw-r--r-- 1 groarchive groarchive    0 mars  20 10:15 tag1.spm
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.spp
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.sps
-rw-r--r-- 1 groarchive groarchive    1 mars  20 10:15 tag1.tmp.tmps

The searchd journal doesn't show anything specificly wrong to me:

mars 20 10:15:02 email searchd[1562]: rotating index: all indexes done
mars 20 10:15:02 email searchd[1562]: rotating index 'tag1': success
mars 20 10:15:02 email searchd[1562]: rotating index 'tag1': started
mars 20 10:15:01 email searchd[1562]: rotating index: all indexes done
mars 20 10:15:01 email searchd[1562]: rotating index 'note1': success
mars 20 10:15:01 email searchd[1562]: rotating index 'note1': started
mars 20 10:15:01 email searchd[1562]: caught SIGHUP (seamless=1, in queue=2)
mars 20 10:15:01 email searchd[1562]: caught SIGHUP (seamless=1, in queue=1)
mars 20 10:05:02 email searchd[1562]: rotating index: all indexes done
mars 20 10:05:02 email searchd[1562]: rotating index 'delta1': success
mars 20 10:05:02 email searchd[1562]: rotating index 'delta1': started
mars 20 10:05:02 email searchd[1562]: caught SIGHUP (seamless=1, in queue=1)
mars 20 10:00:02 email searchd[1562]: rotating index: all indexes done
mars 20 10:00:02 email searchd[1562]: rotating index 'tag1': success
mars 20 10:00:02 email searchd[1562]: rotating index 'tag1': started
mars 20 10:00:02 email searchd[1562]: caught SIGHUP (seamless=1, in queue=1)
mars 20 10:00:02 email searchd[1562]: rotating index: all indexes done
mars 20 10:00:02 email searchd[1562]: rotating index 'note1': success
mars 20 10:00:02 email searchd[1562]: rotating index 'note1': started
mars 20 10:00:02 email searchd[1562]: caught SIGHUP (seamless=1, in queue=1)
mars 20 09:45:03 email searchd[1562]: rotating index: all indexes done
mars 20 09:45:03 email searchd[1562]: rotating index 'note1': success
mars 20 09:45:03 email searchd[1562]: rotating index 'note1': started
mars 20 09:45:02 email searchd[1562]: rotating index: all indexes done
mars 20 09:45:02 email searchd[1562]: rotating index 'tag1': success
mars 20 09:45:02 email searchd[1562]: rotating index 'tag1': started
mars 20 09:45:02 email searchd[1562]: caught SIGHUP (seamless=1, in queue=2)
mars 20 09:45:02 email searchd[1562]: caught SIGHUP (seamless=1, in queue=1)
mars 20 09:35:02 email searchd[1562]: rotating index: all indexes done
`` 

I've launched a manual reindex via `sudo -u groarchive reindex -ap`.
After an hour or so, I got `put 74274 messages to sph_index table for reindexing`
Still no other mails than today's shown in grommunio-archive and the `/var/lib/grommunio-archive/sphinx/main,*` files are still zero bytes.

I've also changed `/etc/grommunio-arvhive/config-site.php` to add specific indexes to sphinx by adding the following line and restart grommunio-archive service.

$config['SPHINX_MAIN_INDEX'] = 'main1,main2,dailydelta1,delta1,tag1,note1';


I'm all out of ideas. Anyone has some ?
Thank you.

Using grommunio-archive-1.3.13.g137.d1b0df1b-lp156.9.166.x86_64

ozy

Btw, some SQL queries show that mails "exist":

MariaDB [groarchive]> SELECT COUNT(*) FROM v_messages;
+----------+
| COUNT(*) |
+----------+
|   83711 |
+----------+
1 row in set (0,560 sec)

The following query also shows recently received / sent emails.

MariaDB [groarchive]> SELECT * FROM v_messages ORDER BY id DESC LIMIT 10;7

But these queries seem a bit off

MariaDB [groarchive]> SELECT COUNT(*) FROM sph_index;
+----------+
| COUNT(*) |
+----------+
|       16 |
+----------+
1 row in set (0,000 sec)

MariaDB [groarchive]> SELECT COUNT(*) FROM sph_counter;
+----------+
| COUNT(*) |
+----------+
|        1 |
+----------+
1 row in set (0,000 sec)

ozy

I'll answer for myself. Main index was broken for whatever reason.
Here's what I did to resolve the issue:

systemctl stop searchd
systemctl stop grommunio-archive
cd /var/lib/grommunio-archive
mv sphinx sphinx.old
mkdir sphinx
chown groarchive:groarchive sphinx
chmod 750 sphinx
sudo -u groarchive indexer --all

Now wait for the indexes to be created. Once done, you can start the services and fix fully reindex the archived files:

systemctl start grommunio-archive
systemctl start searchd
sudo -u groarchive reindex -c /e-tc/grommunio-archive/grommunio-archive.conf -ap

Hope this will help someone.

PF_JHB

ozy

It works 🙂

Thanks