The answer in short to spare you any misleading in this case:

It's three known bugs: In sqlite, and/or in Firefox's handling of sqlite, and in the operating system Win 7. (Win 7 shouldn't allow a program to stall the whole system.) Deleting the faultily huge "places.sqlite-wal" in the Firefox-user's profile folder solved it.

My notebook system is: Win 7 Home Premium 64 bit. One classical harddisc and one SSD are in the notebook. The SSD is the boot disc. Intel processor with two cores with two threads each. (4 threads together.)

Right now seconds but mostly 5 to 20 minutes after startup my system stalls.

When the system stalls, then the task manager (if still accessible) shows that "NT Kernel & System" (always with PID 4) goes up to 25 and 26 % cpu usage. The program window I was working with when the stalling started is partly responsive, sometimes not responsive at all, sometimes when clicking on the window it then gets covered with white half opaque.

Sometimes others program windows are still full accessible. Sometimes ctrl+alt+del takes me to the win user logof/change/ etc page, but though there sometimes clicking is possible - it never does work whatever I select, if anything it takes me back to the desktop and program windows.

After start of the stalling it takes seconds up to 2 minutes and then nothing of the system is accessible anymore.

Only holding the on/off button for seven seconds works anymore to turn the notebook of by stopping the power.

I tried process explorer already, with process priority setting to maximum. But once the stalling appears, I can't save anything from there anymore. (Could I make a continually saving log with process explorer?) I also tried WhySoSlow beta, but nearly to the same avail. I didn't try to record everything (processes) with WhySoSlow so far, because I have bad experiences with full steadily running logs. (But I will record everything if neccessary.)

My question: What should I try now?

I'm really out of ideas, as even Process explorer on highest process priority stalls.

My guess is that the Win 7 system has acquired a malfunction (or changed setting) so it stalls when the HDD is going to sleep or is woken up again. Running music from the HDD in an endless loop I had no stall for hours. But when writing the last sentence of the first version of the long text I already had written for here, it stalled again. ): Maybe the whole mp4-music file ended in cache and so Win 7 again tried to put the HDD to sleep? (Or wake it up later on.)

The system worked fine for at least over 6 months. (Actually I believe I put the SSD over two years ago into the computer.)

When writing a necessary longer letter around 28 December 2015 the system started with stalling. Beside other things I tried: I exchanged the classical harddisc by another hdd. And put the two discs (new HDD and SSD) in another technically identical notebook. The stalling happened there, too.

So I already narrowed the cause: 1. It's either the SSD. or 2. My Windows 7 has acquired some problem it didn't have before. (I can't exclude that I changed some energy settings, to make the Notebook more silent, for example by letting the HDD going sooner to sleep.)

I now tried to run an mp4-file (located on the HDD) in an endless loop, and the system is running now for hours!

So I assume, that it has something to do with how Win 7 handles the HDD or the SSD.

--- Oh no! Just when I was writing the last sentence the system stalled again after hours. ): I changed the wording of the question, and now ask in general.

Please help!

(P.S. I have a 64 kib/sec internet connection currently.)

Notebook is:

MEDION P7624, Microsoft Windows 7 x64 Home Premium, Intel(R) Core(TM) i3-2350M CPU @ 2.30GHz, 4.0 GiB expandable to 32.0 GiB

  • have you added a HDD caddy to use the 2 drives? If yes, look if the caddy has a jumper and change that jumper position. – magicandre1981 Jan 4 '16 at 5:26
  • capture a xperf trace of the high CPU usage of the DPCs: forum.sysinternals.com/… – magicandre1981 Jan 4 '16 at 18:08
  • @magicandre1981: I would already have done that. But as I wrote I currently only have 64kib/sec. ): And I'm pretty sure I have found the reason (or the initial cause) for the stalling. The comp is running now fine for some hours. If it doesn't stall for 2 or 3 days, I'll put the reason into my question text. (Though it might be a bug that could be exploitable, and then I better should post it on the causing software's bugtracker site first.) – John Jan 5 '16 at 5:51
  • put your steps how you solved it in an answer, do not add them to the question, – magicandre1981 Jan 5 '16 at 16:59

The solution:

There are occurences of this reported for over 3 years on the mozilla-bug website.

It's Firefox stalling the system when accessing too often or with too big data sizes the file "places.sqlite-wal" in the Firefox-user's profile directory. Related or as cause of this that file becomes very large. My "places.sqlite-wal" became over 1.8 GByte.

After deleting the too huge "places.sqlite-wal"-file no more stalling occured.

[ You might increase the sessionstore-time in about:config, but I do not know if this advice I read somewhere really is a solution. At least I can't see why making a session store every 15 seconds default (value 15000 ms) is necessary.]

In the internet I found examples with Firefox running on Win 7 Home Premium and using an SSD or an classical HDD. The biggest file size reported I found was 6.5 GByte.

I class this as three serious bugs:

  1. Firefox seems to keep allegedly deleted info in the "places.sqlite-wal" file, as sqlite itself or Firefox's handling of it seems to be faulty. A BIG SECURITY AND PRIVACY BREACH!
  2. Win 7 should definitly stop Firefox from being able to make the whole system 100% unresponsible for a long time. I found cases from 90 seconds to cases where the system still didn't respond after 10 minutes.
  3. The sqlite-website itself reports about errors happening when writing too often or too much to or in a too large wal-file.

I go even further and say that Firefox's switch to the use of the sqlite-database was fishy and it's handling of bugs related to this sqlite is fishy: Mozilla's staff comments on that topic were not helpful, to put it mildly. The problem is called as seldom.

In my case it was only possible by chance to find the reason, as ALL process monitoring programs I tried were not able to record what was happening - even Process Explorer with thread priority set to the maximum didn't respond anymore. (I wasn't able to download Microsoft's XPerf suite due to my slow internet connection with 64 kbit/sec.) People will in the end just make a new installation of Firefox or even the OS instead of reporting the bug.

My own conclusions after reading those things I found was that the sqlite use allows spying on the user.

Idependently when writing about the big file size someone else told me that the places-files are the most yielding ones in computer forensic.

P.S. How I found it? By chance. As last resort I was about to prepare cloning my boot partition to another harddisk. Because I only had about 35 GByte on the other harddisk I tried to delete huge files from my boot partition. Having an SSD it was quite riskless to run a search in Windows Explorer over the whole c-partition for the biggest files, and "places.sqlite-wal" was the biggest. (I already removed win7's and x-systems' installation files and 3 movies.) "places.sqlite-wal" was by far the largest file then. After deleting it, the stalling never occured again.

All the other things I did were lost time.

Some links: https://bugzilla.mozilla.org/show_bug.cgi?id=871908 And explicitly about file system problems somewhere in http://www.sqlite.org/draft/wal.html

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.