Closed dncrews closed 1 year ago
The largest user export I've tested is about 2-3 million users. It might help to break it up into chunks, eg, alphabetically (A-N, O-Z), or by individual letter.
You could also script it yourself using JS/Python/PowerShell, etc. I have example code for this.
I helped another customer with a similar issue (via email). Let me find it and recap it for you.
In terms of Okta logging you out or computer going to sleep, I'm afraid I can't help you. I could (in theory) write smaller CSVs, but how would you resume?
I should add, the size of the CSV definitely has a cap due to Chrome (and Firefox's cap is larger). So, it also depends on how many columns you're selecting/saving to the file.
from Apr 2022 email with another customer who has 2 million users:
it appears that the Chrome limit is 512 MB... Firefox doesn't have the 512 MB limit
rockstar has a byte counter. if you keep your eye on it and the user count, and compare it to 512 MB, you'll know when you're about to reach the limit.
search:
search=profile.lastName gt "E" and profile.lastName lt "G"
Okta has a built-in report called Okta Password Health (see Reports > Reports). it wil show you the following fields:
User,Login,Status,Activation Date,Authentication Source,Last Login,Last Password Change
The User field has first/last names. This might help you figure out the alpha buckets.
Note, I'm not sure if Okta's report will work with 4 million users. If it doesn't, please let Okta Support know.
rockstar has a byte counter. if you keep your eye on it and the user count, and compare it to 512 MB, you'll know when you're about to reach the limit.
Yeah, that'll do it lol.
Thanks. This is a lot of really good information. I'll see what I can do to break this down further, either by name or creation date. I can even try Firefox for the simplicity of doing it once, though I'm absolutely dreading they might have some other limit that I may hit.
I do wish if I were getting close I could say "ok that's good stop now and export it", but that's not a deal-breaker.
... "deal-maker"? Either way.
any news?
i tried an experiment where i opened 4 browser tabs and had each one download 1/4 of the users (eg, by first letter of last name). it seems to work ok, and it's about 4 times faster. this should scale to larger numbers of tabs, too (exercise left to the reader).
i'm going to close this for now. feel free to reopen if u have any updates or questions.
this should be fixed by https://github.com/gabrielsroka/gabrielsroka.github.io/pull/57
i just submitted it to Google Chrome Store for approval
thanks again to @stashev
I spent several hours yesterday trying to download 4MM+ users. As you can expect, this process took 3+ hours, and it failed a few different times. In each failure, I wished I could get a copy of the partial results I'd already generated, and I wished I had the ability to resume from that point.
Uncaught RangeError: Invalid string length
. Apparently 4MM is too many to be able to export.If there is a way to do this already, that works for me as well, and I know this isn't likely a priority for most. I just wanted to bring it up.