Jump to content

Trace Windows 7 boot/shutdown/hibernate/standby/resume issues


MagicAndre1981

Recommended Posts

I've uninstalled Avast (and Sandboxie as I need to reinstall it) and I also disabled something in the BIOS called "SATA IDE Combined Mode" which seems to have fixed the BSOD I was getting when running Plex Media Server. It also installed two new devices/drivers for my IDE ports, so I think before they were somehow connected to the SATA controller and should now be separate.

Anyway, even though my BootPostBootTime has come down from around 80s to 43s, my MainPathBootTime has now doubled from around 22s to 42s, giving a current boot time of around 85s.

I see the ExplorerInit stage now takes about 24s whereas before it was only 3.8s!

The drive enumeration time has come down a bit (4+X+0+0 seems to have disappeared). The SATA drive connected to the AHCI controller still shows as an IDE device here, so I guess that's just a quirk of the report. This is with my IDE DVD connected, I'll try adding my IDE HDD as well next.

<phase name="bootStart" startTime="39" endTime="5926" duration="5887">

<pnpObject name="PCIIDE\IDEChannel\5+1270fc08+0+0" type="Device" activity="Enum"

startTime="1393" endTime="4906" duration="3513" prePendTime="0" description="IDE Channel" friendlyName="ATA Channel 0"/>

<pnpObject name="PCIIDE\IDEChannel\5+1270fc08+0+1" type="Device" activity="Enum"

startTime="1394" endTime="4906" duration="3512" prePendTime="0" description="IDE Channel" friendlyName="ATA Channel 1"/>

<pnpObject name="PCIIDE\IDEChannel\4+223a9b9c+0+1" type="Device" activity="Enum"

startTime="253" endTime="1267" duration="1014" prePendTime="0" description="IDE Channel" friendlyName="ATA Channel 1"/>

<pnpObject name="IDE\DiskSAMSUNG_HD103SJ_________________________1AJ10001\5+2ba897a2+0+0.0.0" type="Device" activity="Start"

startTime="4933" endTime="5877" duration="944" prePendTime="944" description="Disk drive" friendlyName="SAMSUNG HD103SJ ATA Device"/>

TVService and MPExtended still show a Total time of around 4.5s each but that's something I'll have to put up with unless the authors can reduce it somehow.

I tried looking at the Disk Graph but didn't really understand it and couldn't see the Queue Depth, so I'm not sure if that's improved or not.

ETL at https://docs.google.com/file/d/0B1fDI89phEESTHBRYjJ5UV9SQ2s/edit?usp=sharing

02_summary_end.xml

Link to comment
Share on other sites


have you run the optimization? Changing an AV tool is a huge change and now you should train Windows a bit.

OK, I'll try that, if I can get the NIC working again that is as it's decided to go funny and says there's no network cable connected. I had the same problem before whenever I had my CF Card IDE adapter connected but I've tried disconnecting any IDE devices and it's still not working, so I might need to get yet another flippin motherboard :angry:

Link to comment
Share on other sites

Hi, I've been reading through this thread after having run

xbootmgr -trace boot -traceFlags BASE+CSWITCH+POWER -resultPath C:\TEMP

with what I thought was success... At least, I didn't have any errors in the process, until I ran the first xperf trace on the ETL file, and got the annoying and cryptic message:

applying restriction of access for trace processing
113409 Events were lost in this trace. Data may be unreliable.
This is usually caused by insufficient disk bandwidth for ETW logging.
Please try increasing the minimum and maximum number of buffers and/or
the buffer size. Doubling these values would be a good first attempt.
Please note, though, that this action increases the amount of memory
reserved for ETW buffers, increasing memory pressure on your scenario.
See "xperf -help start" for the associated command line options.

I tried adding the various switches to the line (-BufferSize 512 -MinBuffers 128 -MaxBuffers 1024), etc. but nothing made a difference, I'd always get the same error.

The ETL is only 25 MB, and I've uploaded it to Dropbox should you have a chance to look at it. I'd sure like to get past the xperf issue, though. My hard drive is a 1.5TB decent Hitachi, with several partitions, but there's lots of empty space. And I have 8GB of RAM and a pagefile that is also 8GB.

I did run ProcMon and turn on boot logging for another go round, but I couldn't see what to do with all the 18 pml files it created, adding up to 4.5GB!

Any and all suggestions welcome.

Thanks,

Here's the link to my boot_BASE+CSWITCH+POWER_1.etl

Link to comment
Share on other sites

my last post is gone (for what ever reason).

The trace is corrupted/incomplete and you should run a new boot trace.

I ran a new boot trace with xbootmgr, and the file came out to be 200MB, better than before, but still, I cannot run xperf on it, and when I open it in Performance Analyzer, it warns that over 43,000 events are missing.

Errr... wait... looking backwards in the posts today, I see that mine is gone, too... dang! So here goes... rewriting...

Between the time of my first post and this morning when you responded, I discovered the new Win Performance Toolkit 5.0. Using the new Windows Performance Recorder, I ran a successful round of 3 On/Off: Boot Traces, and opened the resulting ETL files in the new Windows Performance Analyzer. I took screenshots of two graphs that I thought are most helpful, they are below. Then I went ahead and ran the "xperf /tti" command against one of the 3 ETL files, and I have attached the summary.xml. It seems like it extracted the right results, so for boot traces, I will just use the new tool from now on. I'd like to find out if the new Recorder tool can also do the Prefetch/ReadyBoot defrag cycle...

One more thing, I just discovered a page in German where someone described how to actually fix the buffers/memory issue. Page is at http://www.wintotal.de/windows-performance-toolkit-macht-mudes-windows-wieder-munter/ and basically he says to change the regkey HKLM\SYSTEM\CurrentControlSet\Control\WMI\Autologger\ReadyBoot\MaxFileSize from 20 (decimal) to 60 (decimal), and a couple other useful things I hadn't known about... so after all this, maybe I will give xbootmgr one more try at tracing.

This is the Boot Phases graph:

7uaw.png

... and this is a mocked up shot of the Lifetime by Process graph, well, a thumbnail bcuz it's too obnoxious & large an image to just embed.

onr2.th.png

Thank you for any suggestions you can provide as to what I can do to better my boot time. When I started working on this last week, Restart time was 133 seconds, so it's down to about 68, pretty good already!

>>>>8 hours later... OK... I have changed the buffer/memory settings and this is what fixed my issues with losing events and with not being able to run the xperf analysis of the xbootmgr etl files. It was so simple!

So, I have run a completely new boot trace, and attached the summary xml file. As of this update, I am deleting the previously mentioned version, and am including the latest one which came from the successful run of xperf.

It looks like my there two major slowdowns on my system: Avira and Windows Search... I'm considering dumping the built in search if I can find a good alternative that includes content searches (maybe dnGrep...) but not sure what I can do about Avira, as I'm a beta tester for tehm and can't (won't) use another AV.

Thanks,

summary_boot.xml

Edited by schmohawk
Link to comment
Share on other sites

have you run the optimization? Changing an AV tool is a huge change and now you should train Windows a bit.

Thanks. Doing that reduced it from 61s to 40s :)

I've disabled most of my boot apps and don't even have an AV or Firewall installed at the moment whilst I debug some issues, so it will increase once I re-add those but hopefully I can keep it down to something reasonable (originally it was something like 266s with all the apps loading!).

Link to comment
Share on other sites

Is it safe to use

xbootmgr -trace shutdown -noPrepReboot -traceFlags BASE+CSWITCH+DRIVERS+POWER -resultPath C:\TEMP

for shutdown tracing or do we need to leave DRIVERS out, as we do for boot tracing to avoid BSODs?

Link to comment
Share on other sites

Hmm, my boot time has crept up again, to 108s, then 112s and now down to 85s, whereas it was as low as 44s before.

I've changed the SATA ODD cable and the IDE HDD cable as they turned out to be dodgy and put the PCI tuner card back in but otherwise I don't think I've changed anything.

I'm not seeing any troublesome durations on the storage devices like there were before.

The only degradation issues listed in the logs are

This application took longer than usual to start up, resulting in a performance degradation in the system startup process:
File Name : lpuninstall.exe
Friendly Name : LastPass Installer
Version : 2.0.20
Total Time : 7177ms
Degradation Time : 2177ms
Incident Time (UTC) : ‎2013‎-‎07‎-‎16T16:06:17.749600400Z

This application took longer than usual to start up, resulting in a performance degradation in the system startup process:
File Name : svchost.exe
Friendly Name : Host Process for Windows Services
Version : 6.1.7600.16385 (win7_rtm.090713-1255)
Total Time : 6450ms
Degradation Time : 1450ms
Incident Time (UTC) : ‎2013‎-‎07‎-‎16T16:06:17.749600400Z

which don't explain the increase.

etl uploaded to https://docs.google.com/file/d/0B1fDI89phEESZHAyYkNqX0dBNTA/edit?usp=sharing

Link to comment
Share on other sites

the TVservice starts slowly.

Most of the delay is caused by all tools you run at startup (47s):

<interval name="PostExplorerPeriod" startTime="39836" endTime="86536" duration="46700">
And again your HDD is busy during boot which is the bottleneck for you.
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...