For those like me who experienced withdrawal yesterday evening, we apologize for the unannounced forum interruption.
The hosting company we use physically moved the servers to a new datacenter a number of miles from their old location. It had been planned for some time back, but that move was scrubbed. They'd sent an advisory heads' up for this one as well, but didn't confirm the actual shutdown before it happened. It took Steve and me by surprise since we didn't find out until after the forum was down.
It was amazing how quickly they got everything back online and we appreciate everyone's understanding.
Michelle
I'd been gone for 5 days and thought the issue was with my computer as it had updated several times while I was away. Good to know all is well.
All's well that ends well.
Would this be related to last nights interruption?
Trying to use the search function and this is what I get.
That's just a sample.
What were the search terms? (Steve says if you are embarrassed about what you were searching for, you can PM him ;D )
If you were trying to search for too short a string, it might get confused and do that.
I definitely had withdrawals last night, but was so happy, when it worked this morning.
HAHA! I was trying to find out what the correlation was between the AquaHot temperature and the ambient temperature.
Years ago there was a post about that, but couldn't find it. Couldn't find anything on WIKI either.
I would like to know if anybody knows the difference between a Gazinta and a Watchamacallit?
Is it working now? Might there have been a "special character" like an "&" in the search terms?
Yes a gazinta is very similar to a guzunder
I just thought it was on my side as our internet provider is pretty squirrelly sometimes. When the Forum was back live this morning , I was just pleased enough to write a silly response to a previously started thread.
Thank you Steve and Michelle for giving us all a helpful outlet for the use of our technology.
Michael
a bit unrelated, but i prefer to use google to search the forum. just enter your search terms and add site:foreforums.com - it will only return results from the forum
aquahot coolant site:foreforums.com
Have you tried the search under
Community -> Search ?
You can specify relevant, most recent, oldest. Search only some boards or all, and add "posted by member" in the event you are looking for a specific member's posts. That's what I typically use.
ETA - direct link here https://www.foreforums.com/index.php?action=search
Nope, never looked that far. I used the search at the top of the page. I'll try the search under Community next time, thanks!
I'm having problems using the search function. Used it yesterday with no problems. Today I get this message. " Unable to access the search daemon " does anyone know what my problem is?
Steve's been working on upgrading the search tool this morning - that might be the cause.
Update complete - try it now.
All is good, thank you!
Apologies for todays Search Outage
Several questions / observations about search had come up, prompting me to investigate some available updates.
We were running Sphinx (https://sphinxsearch.com), specifically version 2.2.11—the last stable release of the V2 branch. When attempting to upgrade to the latest version (3.7.1), I encountered some problems: the search engine failed to start. I tried multiple other 3.x releases, but all produced the same result.
The motivation behind upgrading included:
- Improved morphology support
- Compatibility with newer lemmatization libraries
- Faster and more memory-efficient indexing
- Bug fixes (though software updates often mean trading old bugs for new ones)
- Enhanced distributed search capabilities
Trying to find a solution, I installed Manticore Search (https://manticoresearch.com), a fork of Sphinx. Since it shares the same base, migrating was mostly seamless as the existing config files were compatible. Manticore not only retains everything that version 3.7 promised, but offers additional enhancements.
Some (Hopefully) Improvements
Previously, our search relied on stemmers, which reduce words to their stems by stripping or replacing suffixes. This approach sometimes produces non-words, for example, "running" becomes "run," but "business" turns into "busi", and "octopi" remains unchanged.
With the upgrade, we now use lemmatizers, which identify a word's base form or lemma. This ensures better search accuracy:
- "running" correctly reduces to "run"
- "octopi" converts to "octopus"
- Words with multiple possible roots—such as "dove" (which could be the past tense of "dive" or a noun referring to a bird)—are handled correctly.
Lemmatization enhances search relevancy by preventing meaningless stems and instead produces valid terms that align with the search intent.
Another enhancement is the phonetic search algorithm. Previously, our system relied on MySQL's Soundex implementation, which had major limitations. Now, we use a double Metaphone algorithm, allowing searches to account for words with similar pronunciations. For example, searching on "fone" will also return results for "phone." That helps those of us (myself included) who struggle with spelling or maybe its just my poor typing skills.
That is probably more than anyone wanted to know. I'm continuing to read up on the new engine to see if there are other features we can implement. Be on the lookout for any odd issues when using search.
One other item, when you see your search results words that were part of the search term are highlighted. This is a browser feature to show "hits" but those "hits" are not coming directly from the search engine. The engine responds with a 1-N list of messages/posts that are sorted by relevancy (and a lot of calculations go into that, if you want to know on that just say so) and to jazz up the results words are simply found highlighted.
I don't think I understood a single word of that.
But, Thank You for all the effort you expended to effect the desired result AND the explanation!!
Well, I need to know what Morphology and Lemmatization mean!🤪🤪
The definitions are far worse that those words!!
For the sake of what changed, I'll provide a useless example!! Let's say you have this post:
Original → "The motorhomes were traveling across the country, and many campers were enjoying the views."
Stemming → "The motorhom were travel across the countri, and mani camp were enjoy the view."
Lemmatization → "The motorhome were travel across the country, and many campers were enjoy the view."
With Lemmatization the meaning is intact, and all words are reduced to valid dictionary forms. This search happens in addition to the search terms entered and provides more relevant results.
What he said .....