| Re: USENET sync breaks and goes into endless cycle when it encounters attachment [message #158322 is a reply to message #158321] |
Wed, 18 February 2009 07:39   |
kerryg
Messages: 157
Registered: September 2008
Karma: 0
|
Senior Member |
|
|
Hi Frank. With the help of tech support at the server we ran through the checklists; the tables are fine, they reckon the db shouldn't need resetting, the only errors visible alternate between the two errors I've previously posted; I've bumped up the memory and timeouts in php.ini to pretty high levels, checked consistency a million times, compacted messages, rebuilt themes, nothing seems to be affecting it. Here's one slightly more detailed error message that happened as I was rebuilding the search index (I think it was successful but it's hard to tell as it didn't give me a confirmation) -
Wed Feb 18 7:15:51 GMT 2009 (/home/kerrygal/public_html/ParisForums/index.php:75
/home/kerrygal/public_html/ParisForums/index.php:135
/home/kerrygal/public_html/ParisForums/theme/default/selmsg.php:727
/home/kerrygal/public_html/ParisForums/index.php:1188
) 1053: Server shutdown in progress
Query: SELECT count(*) FROM fud26_msg m INNER JOIN fud26_thread t ON m.thread_id=t.id INNER JOIN fud26_forum f ON t.forum_id=f.id INNER JOIN fud26_cat c ON f.cat_id=c.id INNER JOIN fud26_group_cache g1 ON g1.user_id=0 AND g1.resource_id=f.id LEFT JOIN fud26_mod mm ON mm.forum_id=f.id AND mm.user_id=0 WHERE m.apr=1 AND f.id=21 AND t.id=6348 AND t.replies=0 AND (mm.id IS NOT NULL OR (g1.group_cache_opt & 2) > 0)
_GET: array ( 't' => 'selmsg', 'frm_id' => '21', 'th' => '6348', 'reply_count' => '0', )
Server Version:
Since one single post is breaking the sync of an entire forum and rendering it null and void - is there any way to get an emergency "killswitch" somewhere, a way to say "just skip this post if it's a problem, increment one and carry on"?
[Updated on: Wed, 18 February 2009 07:51] Report message to a moderator |
|
|
|
|
|
|
|
|
|
| Re: USENET sync breaks and goes into endless cycle when it encounters attachment [message #158359 is a reply to message #158343] |
Thu, 19 February 2009 20:52 |
kerryg
Messages: 157
Registered: September 2008
Karma: 0
|
Senior Member |
|
|
Followup: I've created new subforums and started a fresh sync. Two out of three previously broken groups synced fast and flawlessly again, and the third (the biggie with 47,000 posts) is still bringing them across but appears to be working just fine (I'll update this when it's done), flying through previous "problem posts" with no errors. The MIME stuff junks the posts up a lot (as does the *ridiculous* level of >>> quotes this group can get up to) but we'll worry about that in the future; as long as the forums sync, I know I can re-sync later if the MIME encoding issue is tackled and clean it all up then.
I've got max posts = 100 per sync. At that speed I get the occasional "server has gone away" error between every say three to seven syncs, but importantly they're now "benign". I can tell in advance when it's a benign error because it refers to the first post of a given block of 100 posts being transferred - those errors cause no cycling problems, sync just continues as the db catches up.
Being able to maintain a speed of 100 per sync (one sync per minute, 6000 posts per hour) is *very* good - at that speed that's still eight solid hours of syncing to get the posts across. But I'm starting to feel like I can take my eyes off it and trust it.
[Updated on: Thu, 19 February 2009 20:53] Report message to a moderator |
|
|
|
Calendar
Search
Help
Members
Register
Login
Home
]