Help. Deleted AOD scheduler via Repair

💬 General Discussion
1

I did Repair Schedulers and I lost the AOD index. I did it because it wasn’t adding new records to the index and figured it needed to be repaired.

How do I get this functionality back? There are no docs on it that I can find.

2

just create a new scheduler and select ‘index unindexed documents’ from the dropdown, then repeat for ‘Optimise Advanced OpenDiscovery Index’

3

Thanks. I did what you said.

It looks like the cron job has to run before a new Lead gets indexed. My server only lets me run the cron job once an hour. So I’ll have to wait and see if this is working again.

Maybe I’m wrong but AOD seems like more trouble than it is worth. What was wrong with the Standard Sugar Search?

4

If want to disable AOD search, you can do so in admin->AOD Settings and you can just use the basic search

5

Yes, I saw that. In fact I think disabled is the default. I’ll probably just use the Sugar standard search unless I can find a reason for AOD.

While on the subject. when a scheduler says it will run every minute, is that only when SuiteCRM is running or does does the job run even when the Suite is not?

6

if the cron on your sever is running and setup to run every minute then it will run every minuite

7

What is the relationship of the “cron” settings on the scheduler to the cron on the server?

What does cron.php (in Suite directory) actually do? I looked at the code but it didn’t make any sense.

8

Hi Dev, Think of it this way:

The crontab on the server needs to run every minute. It runs cron.php, which basically says “See if anything in the SuiteCRM scheduled jobs list is ready to run. If so, run it. Most of the time, there is nothing to do.” It is totally independent of any other programs running. It just runs the job at the frequency specified.

The scheduled jobs in SuiteCRM are the individual jobs that need to be run and the frequency. The jobs are purely related to SuiteCRM.

======================

Many software packages have all jobs scheduled in the server based Cron. The reason Sugar/Suite does it this way, is so a user can use a web interface to manage the job schedule, without signing onto the server to manage the jobs.

9

Good explanation.

There is a flaw in the Suite logic for this. Many hosts employ a c-panel for users to set cron jobs… and many do not permit users to run a job every minute. Our host only allows once an hour as the shortest frequency. That’s true for many others too, and for good reason… a server with some 100 users all running monster scripts each minute would impact the server.

So if I add a new Lead, I could end up waiting from one to sixty minutes before it gets indexed by Suite, depending on when the cron job last ran.

In a perfect world the Suite logic here would be great, but in the real world, it doesn’t compute.

And why did they even develop it? Where is the advantage over Sugar search?

10

I guess I’m spoiled, my shared host (dreamhost) doesn’t have an issue with running a cronjob every minute. Granted, if I ran a process that uses too much CPU, they’ll talk to me and we’ll fix the problem together. If I ran a VPS, they wouldn’t care as I have full root access.

My guess is that most people running jobs that need to execute frequently (advanced indexing/search, scanning an email every minute to update a case, etc), are running on less restricted servers.

AOD is Salesagility’s implementation of the the Lucene search engine onto Sugar: http://lucene.apache.org/core/
Without getting to detailed, it’s trying to take a google search approach vs. a traditional approach. With the google approach, the database is crawled offline to create a very detailed index that can be searched instantly. The traditional approach is to search the appropriate tables when the user requests the search.

If you want to bring up “John Smith” as a lead, both approaches are fine. If you want to search for “John Smith” anywhere in the database (listed in cases, projects, opportunities, quotes, task notes, etc) the Google/AOD approach is superior. It does take more resources, however.

11

I can see the advantage of this new search. It’s sort of the “index every word” methodology and it works well. Apple’s OS X has similar code in their “Spotlight” search.

I’m talking with my host to see what they will do for me. I’ve been with them a long time.

(BTW, Dreamhost was bought by EIG and you can read on the net what others have to say about what happens to good ISPs once eaten and digested by EIG.)

12

Thanks Dev,

I “Think” the Dreamhost->EIG thing was an april-fools gag a few years ago. I don’t see anything real that they’ve sold-out. Please correct me if I’m wrong.

Full Disclosure: I run our production SuiteCRM instance in-house. I want the best possible performance, and for me, running on Ubuntu on a decent VMWARE box is ideal. I put the HTTPS port into the DMZ so that employees can access it remotely. But as 90% of access is at HQ, I want 0.15 second response times vs 0.75.

13

I think you are right. I had them confused with Bluehost who was bought by EIG.

Running SuiteCRM on an intranet probably works well, but then you have to open it up so users can work offsite… meaning you have to have the same ten tons of security procedures that a professional host/server-farm has. And you have to do your own hardware maintenance. I guess if you have the budget for it, it pencils out.