Web Session Timeouts After Upgrading to 2023.2.1

ragetank, I have been dealing with this web session timeout issue since I originally created this post 4 months ago. I have been back and fourth with SW support trying everything they said, sending diagnostics, rerunning the Configuration Wizard many times - nothing worked. The last ditch suggestion was to build a new Orion server. I did this and followed their DOC to get things going again with out any data loss. New server up and running, licensed and no data lost. Brought up the same dashboard that was giving me the issues before and for about 45 minutes it worked. Then the session timeout again. I was ready to give up until someone linked me your response about changing the database login between the web server and the database. I made this change via running the Configuration Wizard for the millionth time. It is now roughly 2 hours later – NO DROPS !

I can’t thank you enough for this response. It would have been great if the Thwack system had notified me as these responses came in (as I requested), but whatever. The ironic thing is that when my coworker sent me the link to your post, I scrolled up and saw that it was a reply to my post. LOL.

Thanks again. Can’t wait to ask support why they didn’t know this.

All comments

monitoringlife

Which IAM are you using? local, AD, LDAP, SAML? If external is that system enforcing a life expectancy?

Kita

This should be addressed in the 2023.3 release, but we can check the logs to make sure the issue is properly identified. If so, there should be a workaround we can apply. Please PM the case number and we can check.

I have tried AD (our primary), LDAP and even local. All yield the same results.

I sent you the case number. If this is to be addressed in the next update, which is out soon if I remember correctly, then I can wait rather than all of us putting even more work into this issue.

I upgraded to the 2023.3 release. The web session timeout issue remains. I tried a few more things including letting the web session sit on a different dashboard. That made the timeout take about a half hour, but it still timed out.

samp14

Do you have additional web server ? . because we are also facing same issue when we add additional web server behind URL .

No additional web servers. One Orion server - everything running on that server; NPM, SAM and NCM.

Vip3r

did this issue get resolved fo you? i started experiencing the same since upgrade to 2023.2.1. would be good to understand if 2023.3 resolves this but judging from your response it hasnt. i took have a case open and its been weeks with no resolution. we have multiple NOC screens and annoying have to constantly re-authenticate

sguido

We are having the same issue. It seems to happen to everyone, more often to some users. We have three webservers behind a load balancer, and it seems to be impacting all three servers equally. I've asked one of my admins, it seems to happen to her a lot more than the rest of us, to use the webserver on the Orion server to see if there is a difference.

monitoringlife

Also being impacted by this now, despite my account being excluded from timeouts (Disable Session Timeout = yes)

viperxuk

solarwinds support suggested an upgrade to 2023.3 but as someone else posted, this hasnt fixed the issue. multiple tech files have been sent to support and not getting much back on what the issue is and how to resolve it.

Can confirm I encountered this doing a leap from 2020.2.6 > 2023.3. In my case it seem's to only timeout sessions consistently in the NOC view. All other webpages don't exhibit the same behaviour unless in the NOC view. Using the DirectLink account works but i'd rather not have the entire installation open to anyone with a link.

I'm also waiting for Solarwinds Support to review my logs but nothing as of yet, from my brief reading it looks like something is erroring out in the SessionManagerEngine and it dumping the session.

NotQuiteGroot

Has anyone had any luck resolving this issue?

hpstech

This is killing us.

I'm trying to get a new customer on board with additional SolarWinds $ spending but the time outs have many people unhappy.

Someone has to know what's going on in the code, no?

Been back and forth with support and nothing as of yet. I've even turned on the DirectLink account and thats timing out. I'd be curious to see if anyone elses DirectLink acount is experiencing the same issue.

hpstech

Check your event viewer log for 4001 errors. I think I found my issue.

I created a post on this also, I'm experiencing the same issues. Very sporadic session timeouts, can be anywhere from 5-45 minutes.

I may have found the solution to this! I believe the issue is the IIS settings get reset. If you go into your IIS Manager, click on the application pools on the left, then right click on the Solarwinds application and select "set application pool defaults". Then find "Idle Time-out" and set it from 20 minutes to 0. I did that and rebooted my VM running the polling engine and it hasn't timed out in almost 35 minutes!

So it eventually timed out at around 25 minutes of idle time, which is what the web console settings is set to even though the account settings are set to not time out. I just increased it from 25 minutes to 480 minutes, I'll update this if it goes past the previous mark.

Somehow it made it worse. It's definitely an issue with SolarWinds code.

Support found the solution for me, I still can't believe this was the fix but nonetheless.

Try changing your solarwinds server database logon username in the configuration wizard to UPN format. I.e change DOMAIN\user to user@domain

Seems like something with the update doesn't like the legacy format. Let me know if this manages to fix it for you

My Hero, it seems to be working now. Thank you!

vonzoom

Just wanted to chime in and say thank you as well. We've been fighting this issue since at least March (01299411), but after running config wizard on all our nodes and changing over to UPN format we appear to have stabilized. The pages seem snappier too

Glad it helped out! It was an absolute deal breaker for me because we basically had no ability to have a NOC screen logged in. I'd hope they address this in the next patch or atleast make an official KB article for it.

Shooty

Can confirm I find the entire GUI at least 20% quicker to respond and load as well. Fantastic workaround!

Thanks again. Can’t wait to ask support why they didn’t know this.