Hi Roberto, Thanks for the additional information - I'll have to look into my supervisor instance, because it doesn't seem to be operating correctly.
Quick question on the index question from before - in the instance where you had the index assigned to both Invoice and PO data types, is there a way to get Mayan to list out the different document types - for example, under Country A will be two links - PO, Invoices? It would be great if we could treat the doc type much the same way as we treat metadata when creating indexing expressions. Also, somewhat along the same vein, is there a way to automatically set metadata based on document type? For example, if the user selects Pay Slip, a metadata field will be set to "Private" vs. "Public"? Sorry for all the questions! Samir On Tuesday, February 16, 2016 at 2:18:08 PM UTC+8, Roberto Rosario wrote: > > Hi Samir, > > Glad to hear it worked for you! > > The steps you mentioned are indeed needed. Sorry you had to hunt for > these, they were in the documentation in the "Deploying" section ( > http://mayan.readthedocs.org/en/latest/topics/deploying.html#os-bare-metal > ). > > I recommend you use a task monitor like "supervisor" which will launch > uWSGI and Celery upon startup and even restart them if they die. > > If you plan to use only one worker you don't need to specify the queues in > the command line ( > checkouts,mailing,uploads,converter,ocr,tools,indexing,metadata). Not > specifying the queues will cause the worker to service all of the existing > queues. > > If you have scalability issues or a task like OCR is blocking the creation > of thumbnails you can launch another worker (using the exact command line) > and assign it to a single queue and balance out the tasks requests. If you > add other workers you can't use the -B option (Beat) on the second worker, > because only 1 beat generator should run at the same time. You can also run > a separate beat generator, using: > > celery -A mayan beat --settings=mayan.settings.production -l ERROR > > > I recommend this from the start so you don't have to worry about it later > when you start spinning up more workers to handle larger loads. > > On Thursday, February 11, 2016 at 11:59:42 PM UTC-4, Samir wrote: >> >> Thanks Roberto for the prompt response. I've been traveling on business >> and haven't had a chance to look into Mayan until today. Your instructions >> worked perfectly and delivered exactly what I needed. >> >> To answer your question, the two areas of the install where I had issues >> were with Celery and uWSGI. I found I had to run the following command to >> initiate a worker process that generates the beats, because I found I was >> uploading documents but they weren't being processed. >> celery -A mayan worker -l INFO -Q >> checkouts,mailing,uploads,converter,ocr,tools,indexing,metadata -Ofair -B >> >> Also, I found I kept getting a 502 bad gateway error, and when I looked >> into it, it appeared to be caused by the fact that uwsgi was not running, >> so I had to manually kickstart that process with the command: >> uwsgi --ini /usr/share/mayan-edms/uwsgi.ini & >> >> I'm not very good with Unix, so it very well could be caused by some >> incorrect setting on my test server. Thanks again! >> >> Samir >> >> On Saturday, January 30, 2016 at 6:38:37 AM UTC+8, Roberto Rosario wrote: >>> >>> Thanks! >>> >>> The deployment instruction have a step setting up Celery workers. What >>> issue did you have with the Celery instructions? >>> >>> The demo is open ended that Invoice document type and Index were created >>> by users testing the project. >>> >>> This is how I solved your use case: >>> >>> - Create the Country metadata. I used a lookup of 3 countries for >>> this test. >>> - Create two document types: Invoice and Purchase Order. >>> - Assign Country metadata type to both document types. >>> - Create an index for Invoices (all and per country) as pictured. >>> Link this index to the Invoice document type. >>> - Create an index with the same structure for Purchase orders (all >>> and per country) as pictured. Link this index to the Purchase order >>> document type. >>> - Create a 3rd index with the same structure for but link this one >>> to the Purchase order and Invoices document types. >>> >>> >>> On Thursday, January 21, 2016 at 10:58:29 PM UTC-4, Samir wrote: >>>> >>>> First off, thank you for developing this software! Installation was >>>> relatively easy, with the minor exception of getting celery to execute as >>>> I >>>> had to hunt around for instructions on that - if I was confident of what I >>>> was doing, I would contribute the celery instructions to the deployment >>>> docs as they're missing. >>>> >>>> I have a couple of questions on Indexes that I hope somebody can help >>>> with: >>>> >>>> 1. On the demo site there's an Index titled Invoice Number. It is >>>> tied to document type Invoices. Yet, when I upload a document type >>>> Invoices, that Index remains empty. Has that Index been created >>>> incorrectly? >>>> >>>> 2. The problem I have that I do not know how to solve is as follows: >>>> - Imagine I have two document types - Purchase Orders and >>>> Invoices. >>>> - Each document type has the "Country" as its metadata. >>>> - I want to be able to click an Index that shows all Purchase >>>> Orders regardless of Country; another Index that shows all Invoices >>>> regardless of Country; and an Index for each Country (I have no issues >>>> pre-defining the country Indexes) which will show all Purchase Orders and >>>> Invoices for only that Country. >>>> - Is this possible and if so, how do I do something like this? >>>> >>>> Thanks! >>>> >>>> Samir >>>> >>>> >>>> -- --- You received this message because you are subscribed to the Google Groups "Mayan EDMS" group. To unsubscribe from this group and stop receiving emails from it, send an email to mayan-edms+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
