Pancomp60
asked on
Windows 10 Locks Up
I have a computer that when a specific user logs in it will randomly lock up on them. For instance it has locked up while they are using it, and while using different applications. It has locked up during the night after they have logged out (but kept power on). It has locked up while in screen saver mode upon returning from lunch. These lock ups occur maybe once or twice a week. I took the computer out of their building to my office and ran it for 2 weeks and could not duplicate the locking up, even when logged in on their account. Did a complete reinstall of Windows 10 just to be safe and ran for another week with no lockup. After returning it to them it locked up again within 4 days of their using it.
It is a Dell OptiPlex and Dell keeps saying it's a Windows issue. Has anyone come across this?
It is a Dell OptiPlex and Dell keeps saying it's a Windows issue. Has anyone come across this?
Jackie Man
Have you checked the Event viewer?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
I think I saw this once but with a Dell laptop and we ended up issuing a service request for replacing the motherboard if I recall or SSD. The station would freeze everytime they would set it down practically or move it. It could be the PSU as Andrew stated. After we reimaged the station, updated everything from the manufacturer site, drivers, BIOS (firmware) etc., we found it to be a hardware defect as it persisted still. Otherwise, if its locking their actual account rather than a freezing lock behavior you could try https://www.netwrix.com/account_lockout_examiner.html
I have seen employees get back-to-back viruses after a reinstall or replacement too, so that is possible. You already did a reinstall of the OS so that could indicate a hardware or process fault as well. You should run the built-in Dell hardware diagnostics, look at Windows Event Viewer > System, Application etc. log sections for error IDs or issue events as suggested above already. Another good tool is Process Monitor https://docs.microsoft.com/en-us/sysinternals/downloads/procmon sometimes processes cause hiccups too.
I have seen employees get back-to-back viruses after a reinstall or replacement too, so that is possible. You already did a reinstall of the OS so that could indicate a hardware or process fault as well. You should run the built-in Dell hardware diagnostics, look at Windows Event Viewer > System, Application etc. log sections for error IDs or issue events as suggested above already. Another good tool is Process Monitor https://docs.microsoft.com/en-us/sysinternals/downloads/procmon sometimes processes cause hiccups too.
Since you don't get the problem offsite while logged in as the user, and you have eliminated Windows as the likely cause by re-installing it, then I concur that the issue is likely to be environmental or related to the user's actions.
One environmental difference might be a docking station, so if the user is connecting to a docking station but you are not when you test off-site, I would eliminate that difference in your tests.
Although it might be that the user's actions are risky, as some have suggested, it could also be innocuous actions that the user performs but you don't. For example, it could be that the computer locks up every time it wakes up after a job has been printed on a particular printer. Since it sounds like you have mostly left the computer idle while you were testing it off-site, then you won't have simulated this particular combination of events, especially if the printer in question is only available in the user's office.
Perhaps the user could keep a log of their activities so you could see if a particular activity precedes the problem. I would start off at a high level to help with compliance (e.g. 8-10: entered orders; 10-11: wrote report, 11-12: responded to email), and drill down to lower levels if the high level logs don't give any clues. I would also presume that rebooting the computer might prevent the problem from occurring, so you might want to have the user reboot the computer on a regular schedule, say, just before lunch and just when they leave for the day (or perhaps turn the computer off at night and on the next morning). This may reduce the frequency of the problem, but when the problem does occur there will be less activity preceding the problem which might be causing it.
As for specific causes, the only one that I can think of has already been mentioned by Andrew, and that is overheating. I have encountered overheating of video cards when the cooling fan on the card has seized, and overheating of CPUs when the cooling fan on the CPU or power supply has seized or become clogged with dust. Another way CPUs overheat is when the heat sink becomes unglued (so the contact isn't good enough to draw the heat) or even dislodged. Overheating can also be environmental if, say, the user covers the vents. One particular gotcha is that some laptops have their vents in the fold between the keyboard and the screen, and they won't vent properly when the screen is closed.
If the hardware is overheating just a little then it might only seize under load, for example, if the user opens Outlook or Excel or if the Windows File Indexer starts. This latter service could explain why the computer seizes when it is idle, but doesn't explain why it doesn't seize off-site. On the other hand, the indexer may have nothing to index if you don't open Outlook, so there's a scenario of how an action like opening Outlook can make the computer seize hours later.
One environmental difference might be a docking station, so if the user is connecting to a docking station but you are not when you test off-site, I would eliminate that difference in your tests.
Although it might be that the user's actions are risky, as some have suggested, it could also be innocuous actions that the user performs but you don't. For example, it could be that the computer locks up every time it wakes up after a job has been printed on a particular printer. Since it sounds like you have mostly left the computer idle while you were testing it off-site, then you won't have simulated this particular combination of events, especially if the printer in question is only available in the user's office.
Perhaps the user could keep a log of their activities so you could see if a particular activity precedes the problem. I would start off at a high level to help with compliance (e.g. 8-10: entered orders; 10-11: wrote report, 11-12: responded to email), and drill down to lower levels if the high level logs don't give any clues. I would also presume that rebooting the computer might prevent the problem from occurring, so you might want to have the user reboot the computer on a regular schedule, say, just before lunch and just when they leave for the day (or perhaps turn the computer off at night and on the next morning). This may reduce the frequency of the problem, but when the problem does occur there will be less activity preceding the problem which might be causing it.
As for specific causes, the only one that I can think of has already been mentioned by Andrew, and that is overheating. I have encountered overheating of video cards when the cooling fan on the card has seized, and overheating of CPUs when the cooling fan on the CPU or power supply has seized or become clogged with dust. Another way CPUs overheat is when the heat sink becomes unglued (so the contact isn't good enough to draw the heat) or even dislodged. Overheating can also be environmental if, say, the user covers the vents. One particular gotcha is that some laptops have their vents in the fold between the keyboard and the screen, and they won't vent properly when the screen is closed.
If the hardware is overheating just a little then it might only seize under load, for example, if the user opens Outlook or Excel or if the Windows File Indexer starts. This latter service could explain why the computer seizes when it is idle, but doesn't explain why it doesn't seize off-site. On the other hand, the indexer may have nothing to index if you don't open Outlook, so there's a scenario of how an action like opening Outlook can make the computer seize hours later.
Test the power outlet to make sure the hot and neutral aren't reversed. I have seen it done before and can sometimes cause weird power supply issues...
ASKER
Thank you all for your suggestions. I will be revising my testing plan to include the possibility of over-heating. The building this user is located in is on 2 years old but upon talking with several occupants, apparently there are know power issues throughout the building that I was unaware of. I will be trying to get an electrician to inspect the receptacles at the user's workstation as well as seeing about relocating their computer from an under-the-desk stand, to placement on top of their desk. Wish me luck.
Well, an easy solution to power issues is to install a UPS. That will filter out the spikes and other irregularities, as well as provide an opportunity for more graceful shutdowns during power outages. A UPS costs about $150 and you will have to replace the battery every four years or so, for another $100 considering parts and labor. Bonus; the UPS can also be a convenient power bar and protect other equipment in the vicinity.
apparently there are know power issues throughout the building that I was unaware of
Bingo. Apart from overheating, I think that's the most likely cause of the problem here. I'd second Gary's suggestion about powering the machine through a UPS. Although a proper electrical test should show up any problems too. I had that scenario once though be aware that a proper electrical test needs to run over a 12-24 hour period. The guy that did a building I was working in left a device plugged in overnight and it recorded the periodic spikes.
This is putting the cart before the horse, but if you install a UPS then it will probably come with a program that lets you monitor the power. My brand of UPS (APC) has a program that logs every power spike, dip and failure. That same app sends a signal to Windows to shut down during a power failure when there are only x minutes of battery capacity left.
ASKER
Ok, to avoid anymore comments about UPS, all systems run on a UPS of some sort, depending upon the system. This one is no exception.
That doesn't necessarily rule it out. It could be damaged, especially if there are power issues in the building.
Swap the UPS with another and I would also try a different PSU.
Swap the UPS with another and I would also try a different PSU.
ASKER
Turns out it was overheating. Changed the location of the computer so it gets better airflow. Thanks Andrew Leniart
You're very welcome, glad to hear it was such an easy fix :)
Best, Andrew
Best, Andrew