UPDATE (06-19-2003): I’ve seen a lot of web search hits for this page and I can only wonder if other people are having the same problem as described below. None of these measures worked and I finally had to reinstall my OS. So far it’s been a few weeks and the problem hasn’t come back.
When I woke up this morning I fixed some cinnamon rolls for breakfast.
When I logged into my computer, I was greeted to a rapidly-degrading system. The dreaded problem struck again last night at 5:28am:
[code]Event Type: Error
Event Source: Srv
Event Category: None
Event ID: 2019
Date: 5/4/2003
Time: 5:23:12 AM
User: N/A
Computer: FREESIDE
Description:
The server was unable to allocate from the system nonpaged pool because the pool was empty.[/code]
This problem is driving me crazy. I began an investigation into what this could possibly be. Everything I found on the web referred to this problem in NT4 or an early version of Win2k. I did find one obscure reference that the problem may be solved by not allowing the OS to put the USB root hub into power saving mode. I changed this setting.
But I didn’t stop there. I uninstalled the Intel Application Accelerator. I installed a new version of the Intel INF file package. I also installed SP1a (Service Pack 1a) for WinXP. I think I already had this installed, but I wanted to make sure.
On the solutions for this problem under Win2k and WinNT, Microsoft recommended occasionally running chkdsk to free up the paged pool or something. This doesn’t make any sense to me considering chkdsk is a disk checker and shouldn’t have anything to do with system memory. But nonetheless, I created a reoccurring task in the task scheduler that will run chkdsk once every 2 hours.
I have a gut feeling that the problem is related to power saving stuff since it only happens when the computer is idle for an extended period of time. It NEVER happens when I’m using it locally or remotely. Based on this, I’m hoping the USB root hub change does the trick. I’ll be very upset if the bug strikes while I’m away next week.
If all of this doesn’t work, I’ll have no other choice but to reinstall WinXP and start over from scratch.
I backed up everything except for installed applications to my spare 60GB HD and physically removed it from the computer. This is my ‘offsite’ backup solution. It would take way too many CD-R’s to back up 30 gigs of data (music and pictures take up a lot)
I did laundry this morning and went to Cumberland mall. I bought two new tshirts and board shorts from Abercrombie & Fitch. From American Eagle Outfitters I got another pair of board shorts.
After going to the mall, I checked the 4:45 flight today. Not good. About 30 ‘HK’ revenue standbys showed up and there are only 21 unsold seats. I made a command decision that I will NOT be flying to Hawaii today and will have to try for the 8:15am flight tomorrow via SLC.
As soon as I made this decision, I rebooked my Hawaiian Air flight from 8am to 3pm and rescheduled my rental car pickup to 4pm. I emailed Larry letting him know of the change in plans.
I finished book 5 of the Wheel of Time series. The ending was a bit of a letdown. There are way too many lose threads. Perrin is raising an army in Edmonds Field. The White Tower is in shambles with a splinter faction raising an army to attack and take back the tower. Rand is somehow supposed to marry three girls but hasn’t done it yet. I wonder if book 6 will address any of these things.
It was a beautiful sunny day today so I applied sun block, grabbed a book and headed out to the pool. I nearly fell asleep while lying on my stomach. It was a great feeling. I headed back in after a few hours in the sun.
I marinated some chicken in ‘Caribbean jerk’ marinade. I also prepared an onion by cutting it in half and then cutting a ‘cone’ out of the center. In this cone, I stuffed butter and brown sugar and olive oil. Then I wrapped the onions in aluminum foil. I grilled everything up and it came out pretty good.
The last thing I did today was pack my duffle bag. It’s nice going to Hawaii because I can pack for a week in just one carry-on duffle bag & backpack.
Jeff,
Im seeing this error on one of our servers here at SPRINT KC, can you tell me if it was indeed the power save features causing problem?
Hi Joel,
Unfortunately I was unable to determine what the exact cause of the problem was. I tried several different things, including disabling all ‘normal’ power-save settings (like monitor, HD, etc). I also disabled the power-save features for my USB hubs from the hardware manager. I reinstalled SP1a. I installed the latest and greatest Intel INF files. I installed (and then later uninstalled) the Intel Application Accelerator. Nothing at all seemed to fix the problem. Finally I reinstalled WinXP and re-installed pretty much everything I had before (making no hardware changes at all).
To this day I have not had the problem anymore ever since I reinstalled. It drives me crazy not knowing what caused the problem but I’m pleased it finally went away.
I have this problem also on a Windows 2000 Server with 1GB of RAM. Its a Poweredge 2600 (2Ghz Xeon) with a RAID 5 36GB 10K RPM Array.
I noticed it began occuring ONLY AFTER I installed a Belkin UPS unit which came with a USB connection that I had to enable the USB ROOT HUB FOR!
Your report plus this leads me to believe you’re searching in the right direction. I’ll report my findings if I discover anything.
When it happens on our machine, it actually won’t allow me to go to Start–> Shut Down –> Restart Computer — it says I don’t have permissions (as Administrator, no less). I can only soft-restart by hitting ctrl-alt-del and choosing to restart from there. No idea why the latter works but not the former, since they perform the same action to my knowledge (I know it doesn’t make any sense, but I’ve tested this throughly.
Also, when you try to login if it’s “real bad” it seems to take a few tries before letting you in (no, I’m not not mistyping the password, this is a real problem).
I’m as baffled as you are and this is a mission critical server running apps for 30 people to do their jobs day and and day out.
Rebooting seems to fix the problem.
Power outages seem to cause it (but not always), which points back to the Belkin UPS unit.
As an update, I’ve discovered several tidbits of info:
1) There is a Master Browser memory leak fixed in Windows 2000 SP2. I disabled my computer browser for the time being to test. I noticed it was receiving browser info from an NTSERVER that was in “error code”.
2) You can use a gflags.exe utlity to test for non-paged pool memory leaks if the value of allocations increases over time.
3) Make sure your paged pool is dynamically allocated.
Adam we have the same problem on a Windows Server 2003 Small Business server. We also run the Belkin UPS Software and have frequent power outages in the area. In my instance I believe this might be it!!! I’m going to search for updated software/drivers for the Belkin to see if it helps.