Forms loading cached / old data

Derek C · ‎11-28-2022

Hi all,

We've been having an issue where a record (e.g. and INC) will load old / cached data after it's been updated. Typically this happens when you open it from an email link. The form will show data that has since been changed (e.g. State was updated to work in progress, but is still showing New). After you reload the form, it seems to be resolved.

We are on San Diego, and we are using the Next Experience UI.

I can't find anything on the community regarding this, and ServiceNow support seems to be stumped as well which makes me wonder if we are the only ones experiencing it.

The issue is intermittent and not reproducible on demand.

The one way I've been able to (kind of) reproduce it is by opening two separate web browser windows (one incognito) and impersonating another user. Then updating a record, using SNUtils to switch the node, and opening the record in the other window.

Thoughts on what could be causing this? Has anyone else seen this issue, or is it something unique to us?

Thanks!

Derek C · ‎02-28-2023

ServiceNow support did finally provide a work around for this. You will need to put a ticket in with ServiceNow as there is an update set they can provide that updates the record that needs to be updated, you will not be able to update it even as admin.

The name of the update set they provided me was "SNC PRB1588976 Workaround", if you need this from ServiceNow support, it might be useful to ask for it specifically.

Feedback from support below.

Most Probable Cause:
This issue is caused by a caching issue with the Next Experience / Polaris UI service worker, specifically the "polaris-sw-prefetch-iframe" worker.

Workaround:
This issue can be worked around by doing:
0. Logging in as maint
1. go to sys_ux_managed_service_worker.do?sys_id=4f58aeeb6d251ab126be8ae441962762 (managed-service-workers/polaris-sw-prefetch-iframe)
2. Uncheck Active
3. Click update
4. Visit cache.do
5. Close & reopen browser (or close all tabs on the problematic domain)
6. managed-service-workers/polaris-sw-prefetch-iframe service worker should no longer be registered

OR:

1. Load and apply the update set that is attached to this PRB (sys_remote_update_set_87632745773021101db9169bae5a991f.xml) and proceed with steps 4-6 above.

How to confirm the workaround:
1. Open chrome dev tools
2. Click on Application tab
3. Click on Service Workers in left panel
4. Click on root.js in center panel
5. Confirm importScripts does not contain '/uxasset/externals/managed-service-workers/polaris-sw-prefetch-iframe.jsdbx' in list

View solution in original post

Derek C · ‎12-14-2022

Hello @A_5,

Unfortunately no we have not found a solution at this time... We are upgrading to Tokyo in less than a month and are hoping that it is resolved there. If you are experiencing the issue, can you please consider opening a ticket with ServiceNow Support? I think if more instances of this issue are reported, it might get more attention.

We did find several sys_properties that we thought may help:

glide.cacheman.cache_pair_optimizations.enabled

glide.cache.translations.legacy_flush

glide.cache.upgrade.allowed_enabled

But there isn't much documentation on these (if any), it may be worth checking those out.

If we do find a solution, I will be sure to post it here.

Derek C · ‎01-10-2023

Hi @A_5 we have finished the upgrade to Tokyo and it did not resolve the issue unfortunately.

Did you ever find a resolution? Did you end up putting in a ticket with ServiceNow support? If so did they have any more information?.

I still have a ticket open with them, but it's been kind of a dead end so far. I am hoping if more people report the issue then it will get more attention.

Thanks,

Derek

A_5 · ‎01-10-2023

Hi @Derek C I hoped that you will bring good news when I got an email notification 🙂 We also created a ticket to ServiceNow support and they answered that they already have this bug assigned in their queue as Work In Progress. We are going to see what will happen next. Hopefully they will finally get rid of this issue with one of new patches.

Derek C · ‎01-11-2023

Thanks @A_5! I'm glad to hear that they do have a bug assigned for this and that it's a work in progress.

If I do hear anything, I'll be sure to post here. Until then, I guess we wait for the bug to be resolved.

A_5 · ‎12-14-2022

HI @Derek C

Thank you very much for your response. We thought about the same solution. For sure we're going to open a ticket for that and hopefully they will take a closer look at this problem. If we find a solution I also will be sure to post it here.