Gallery "Out of Memory" Errors on Otherwise Stable Workflows

[jrlindem]

Every now and again, I get scheduled workflow failures where the errors are along the lines of:

• removeNullRows (143): Tool #9: Out of Memory (ToolId: 12)
• Out of Memory (ToolId: 7)
• Unable to allocate memory (ToolId: -1)
• Tool #238: Unable to allocate memory for temp files. (ToolId: 12)
• Unable to allocate memory (ToolId: -1)
• removeNullRows (143): Tool #8: Unable to allocate memory for temp files. (ToolId: 12)
• Unable to allocate memory (ToolId: 49)
• Unable to allocate memory for temp files. (ToolId: 49)

These are from two otherwise very stable workflows that I discovered this morning.  Our Gallery is on an AWS instance.  Last time this happened, I think? our admin's had to clear cache/memory on the server and restart it...

What I'm curious is why does this occur, this seeming accumulation of memory clogging "things" and if any of you have experienced this... how do you resolve or avoid in the first place?

Thanks in advance,  -Jay

[KGT]

I would expect that the server may be configured with swap file on C:\ drive or similar and that has not much space available.

Other things of note:

• Do you have Electron Apps running on the server? I would hope that MS Office etc are not actually running there (but maybe installed for writing to Excel etc).
• Alteryx should clean up after every run, but there is always some memory that may not be released instantly. Nothing that a weekly restart shouldn't fix though.
•  

On the error messages:

• (ToolId: -1) is a workflow message rather than a tool message and so it's just run out of space.
• removeNullRows() seems like it might be from the Data Cleansing Tool. I always advise my team that the Data Cleansing tool is great in Development, but I don't want to see it in production. It's memory intensive because it allows so many options. Switch it out for a Multi-field if you can. (unless it's doing many operations already).
• "Unable to allocate memory for temp files" seems like a swap file issue. This is the main one that advises the above...

[jrlindem]

@KGT 

Appreciate the response.  No Electron Apps that i'm aware of.  This is a dedicated AWS instance for Gallery so it should be limited to just that.  As for the Swap File... that's an interesting thought and will explore that avenue.

And yes... I'm using Data Cleanse despite it's shortcomings.  The couple workflows aren't terribly resource heavy, so I got lazy.  Appreciate the guidance and will explore more deeply.  Cheers,  -Jay

[KGT]

Cool, yeah, I probably wrote a bit, but the most likely thing to check is the tempfile settings on the server in Options > Advanced... > System Settings (I think that's the name, I don't have it anymore as I don't have a server anymore). Often it will be set to an OS drive or something that has 2-4GB free, but you definitely want more than that. (Don't need 500GB or 1TB like help might say, but definitely want a bit)

[jrlindem]

No, no, I really appreciate all the ideas.  I'm having our server admins take a look at that setting to see what is being made available.  I also think it's just a matter of having the memory cleared and the server restarted periodically.  The issue seems to occur pretty spaced out and I think it's just incremental bloating of cache that's not being cleared out on it's own, but knowing how much space is available may give us some hints as to how often to clear it out.

-Jay