Sorry for the multiple posts, but to be clear to potential future readers, I mean the issue seems to have gone away after changing the order and reindexing/restarting. Thanks again.
On Aug 24, 10:44 am, Alan <[email protected]> wrote: > Thanks for the info, the issue *seems* to have gone away but I'll keep > an eye on it. > > Oh and also, I realized that I made a mistake in "symptom 3". It's > actually the model's corresponding .spp and .spd files(and not > the .spa file) > > Thanks again! > > On Aug 24, 9:49 am, Pat Allan <[email protected]> wrote: > > > Definitely an odd issue. I'd recommend calling your id field something > > else... > > > indexes :id, :as => :model_id > > > Or something like that, see if that helps, because having two columns > > called the same thing will likely confuse Sphinx (although why that's > > only started to happen now, I've got no idea). > > > Also, if you're copying the index files across, perhaps copy the > > config file across as well? The more I think about it, the more I > > think the confs being in different orders could be the cause - TS > > calculates an offset for each document to ensure there's unique ids, > > and that's based on the number of indexed models, and the order of > > them according to TS. It has recently come up on this list a week or > > so ago, that different orders means different offsets, and so that > > could be what's screwing things up. > > > I'll try to find time to patch things to ensure the ordering is > > consistent (ie: alphabetical), but that might not happen super-quick. > > In the meantime, copy the conf file across as well. > > > -- > > Pat > > > On 24/08/2009, at 5:40 PM, Alan wrote: > > > > I'm running into some very inconsistent behavior, and I'm trying to > > > track down the potential culprit. > > > > The situation is that on multiple machines(linux distros) running the > > > same sphinx binaries/ts plugin, on one machine the index seems to be > > > missing a good chunk of data(this *seems* to have just happened out of > > > no where). It looks to be an issue with the indexer as when I copy > > > over the indices from the good to bad instances, ThinkingSphinx starts > > > to return the right information. The only thing i can spot is that the > > > conf files have models listed in different orders. > > > > somebody pointed out that I was indexing an "id" field, and the conf > > > file has sql that generates output with two id fields because of it. > > > This wouldn't explain why it works on the one machine and not the > > > other, but juuust in case it matters? > > > > Has anyone experienced similar problems give the symptoms I have > > > given? > > > > To recap, here are the things that I think might be potential issues. > > > 1. Is it problematic to index an ID field? > > > 2. Any chance that the order of models in a conf file cause this sort > > > of issue? > > > 3. I also noticed the model.spa file on the "bad" machine seems to be > > > a lot larger. > > > > Note that I'm running the same sphinx binaries/ts plugin, and the > > > second machine that works is working against a dump file from the > > > first machine(that stopped indexing correctly all of a sudden). --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Thinking Sphinx" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/thinking-sphinx?hl=en -~----------~----~----~----~------~----~------~--~---
