Message boards : Number crunching : Data lost when Computer Reboot
Message board moderation

To post messages, you must log in.

AuthorMessage
Hathor
Avatar

Send message
Joined: 17 Sep 04
Posts: 27
Credit: 298,995
RAC: 0
Message 4517 - Posted: 27 Oct 2004, 23:08:31 UTC

Strange thing happened. My computer (Intel 3.2 HT Win XP SP2) crashed and rebooted for no reason that I know of. No warning, no message no virus or other malicious code.

To the point, I still have data in my project folder, however the GUI does not recognize it therefore I have no WU listed to crunch or the compleated ones to report under the work tab.

I re-installed v 4.13 but no joy.


Any Ideas??




<a> [/url]
ID: 4517 · Report as offensive     Reply Quote
STE\/E

Send message
Joined: 2 Sep 04
Posts: 352
Credit: 1,393,150
RAC: 0
Message 4518 - Posted: 28 Oct 2004, 0:07:59 UTC
Last modified: 28 Oct 2004, 0:08:18 UTC

Yes, that has happened to me a few times, especially before I invested in some Backups for all my Computers, I lost all the WU's on all the computers a few times because of Power Blips ... Since I installed the Backups I haven't had a problem with lost WU's ...
ID: 4518 · Report as offensive     Reply Quote
ric

Send message
Joined: 17 Sep 04
Posts: 190
Credit: 649,637
RAC: 0
Message 4519 - Posted: 28 Oct 2004, 0:15:18 UTC - in response to Message 4517.  

> My computer (Intel 3.2 HT Win XP SP2) crashed and
> rebooted for no reason that I know of. No warning, no message no virus or
> other malicious code.

May I ask very friendly and polite, have you overclocket the pc
til the lights flackered and the cpu began to glow ??

I had the same situation here to. --> After a power failure, no oclock :-), in the project folder, work there, but client_state was reduced to initial size.

It's like a book/FAT, having the pages, but no content table.
Work but no worklist.

Reinstalling a new client vervsion, will not put the work back.
In the project folder, there will be some grufti WUs left.

In those case (rarely arrived), I detached the project (on multi project attached hosts) or directly deleted the boinc folder and reinstalled from scratch.
Make a defrag before reinstalling to keep the file system maintenanced.


> Any Ideas??
it's called backup

If it happens, it happens.

IF the USER is doing something, til someting happens, better have a working backup on a other partition/location as the file system.

An other possibilitiy is a electrical hiccup. On a worst case, if your neighbor is making a cup of coffee, your pc is restarting.

watch if it happens frequently (could be also the hard drive)


ID: 4519 · Report as offensive     Reply Quote
ric

Send message
Joined: 17 Sep 04
Posts: 190
Credit: 649,637
RAC: 0
Message 4520 - Posted: 28 Oct 2004, 0:18:07 UTC - in response to Message 4519.  

poorboy was quicker..

@PB (arrived and operational?)
ID: 4520 · Report as offensive     Reply Quote
Hathor
Avatar

Send message
Joined: 17 Sep 04
Posts: 27
Credit: 298,995
RAC: 0
Message 4521 - Posted: 28 Oct 2004, 1:32:40 UTC


No OC'ing.

Did BU before, however unless it is done just after each change in crunching, upload, download, ready to report, etc. Gave client hissy fits that required detach and re-attach.

USER doing nothing but crunching with an always connected line.

All hardware is top notch and have my very own powerline to my house, let the neighbor make a cup of coffee or even warm it up in their Microwave.

I will just let the client run until their is some more work then it can start all over again.
ID: 4521 · Report as offensive     Reply Quote
ric

Send message
Joined: 17 Sep 04
Posts: 190
Credit: 649,637
RAC: 0
Message 4582 - Posted: 29 Oct 2004, 7:57:43 UTC - in response to Message 4521.  

This night,
one member of the rig, must had a bad dream.

Found the m$ client, w2k running, innocent restartet, the boinc gui windows where full of nothing. Only the content of the Message Tab was showing, the client started and the sheduler responded with no sheduler responded.
And some project down lines to.

Same as U, files there (zip) on level file system, but the user xml, this has a new creation time stamp.:----------(
And it becomes very short.

USER did also nothing here. Was sleeping well.
(This client normaly running 2-3 month til MUst down due hoovering)

may I copy your posting?, so I don't have to write it myself:

>I re-installed v 4.13 but no joy.

No bad thingking, no unrespectfull posting, but if there would be a third,
a new membership could be created.

It's only "sad" for LHC, THEY lost about one day of "work" returnend

Now, the feeling of loosing work is refreshed, the clients queue refilled.
Ready for the next!

Have a nice day
ric




ID: 4582 · Report as offensive     Reply Quote
STE\/E

Send message
Joined: 2 Sep 04
Posts: 352
Credit: 1,393,150
RAC: 0
Message 4586 - Posted: 29 Oct 2004, 10:34:47 UTC
Last modified: 29 Oct 2004, 10:35:12 UTC

@PB (arrived and operational?)
===========

Ones up and running ric with no problems, the other ones put together but not running yet. It's in a kind of endless reboot loop right now, just before It gets ready to go into Windows It gives me a message that there is no Video Feed basically and reboots itself.

It was getting late last night & I just went to bed without getting it running. I'll have to check a few things & see if I can't get it straightened out.

ID: 4586 · Report as offensive     Reply Quote
Profile Michael Berger

Send message
Joined: 2 Sep 04
Posts: 126
Credit: 49,653
RAC: 0
Message 4589 - Posted: 29 Oct 2004, 11:04:32 UTC - in response to Message 4518.  
Last modified: 21 Nov 2004, 12:17:13 UTC

ID: 4589 · Report as offensive     Reply Quote
Toby

Send message
Joined: 1 Sep 04
Posts: 137
Credit: 1,691,526
RAC: 8
Message 4592 - Posted: 29 Oct 2004, 12:05:21 UTC

This computer crashed several times before the old motherboard blew out on me. I never lost all my work units although a CPDN work unit did get restarted once. Are you using the FAT or NTFS filesystem? Files on a FAT partition can get completely hosed by an unexpected reboot - especially if they are in use when the crash occurs. Journaling is the answer! :)


--------------------------------------
A member of The Knights Who Say Ni!
My BOINC stats site
ID: 4592 · Report as offensive     Reply Quote
ric

Send message
Joined: 17 Sep 04
Posts: 190
Credit: 649,637
RAC: 0
Message 4593 - Posted: 29 Oct 2004, 12:13:59 UTC - in response to Message 4586.  
Last modified: 29 Oct 2004, 12:14:20 UTC

tks for updating!

>It's in a kind of endless reboot loop right now, just before It gets ready to go into Windows
guess 99,9% software/driver problem, because you HAVE a video signal
the rest 0.1% cooling/hardware/interrupt problem on level grafic card
the rest 0.000000000000000000000000000001% "unknown" (sorry writing on AMD)

>It gives me a message that there is no Video Feed basically and reboots itself.

ugly, but you will find the way to manage.
further goodluck!

It happened to me, on a XP, not passing a "welldone/propper" installation for a network card, that the system is "hanging". In this case, at boot time, by pressing the F8 Key, the start options can be selected. even with a brocken
video card settings, a nativ VGA mode, (in my case without network) can be selected. In takes time util up, even on ugly VGA mode, over the system proprieties, devices or how it's called in your environment, the faultive driver/component can be "disabled" or deleted/removed (redetected next boot)

Aehm. no need to put the companys name, a while ago, on an customer call,
the problem was, the grafic card had 2 plugs for vga cable. After plugin in the primary plug/jack, the software part could be finished....;-)

but I have seen your numbers doubling in just a couple of days..
It looks like there is enough work for every client!!



>It was getting late last night & I just went to bed without getting it running. I'll have to check a few things & see if I can't get it straightened out.
wishing viel Gl�ck/good luck

parts are @U, a new day started and you have the luck, to have a pc with internet connection on your hand to get newest driver

ric
\\.



BTW
The project needs every cpu cycle it can get.


BTW2
I know you will not listen, but don't do the same error like me,
getting (over month) newer, faster cpus, with the thinking behind, to power off the "old" one. What ever old is. Now running both, the older and the "less older", in perfect harmony.

The only thing (I'm) missing is something like a SetiQ

ID: 4593 · Report as offensive     Reply Quote
ric

Send message
Joined: 17 Sep 04
Posts: 190
Credit: 649,637
RAC: 0
Message 4595 - Posted: 29 Oct 2004, 12:55:03 UTC - in response to Message 4592.  

@Toby
have a remembering of 3 times,
the cl of this night, windows 2000 512mb AMD p1400 heavy oclocked to 1422 MHz.
NTFS.
An other was w2k too, NTFS, I didn't take so much attetion, I BELIVE , it was a w2k too. NTFS.

Do yo know th AI from asus? oclock my mouseclick (great).

one day, p4 512 MB HPFS, after backup on other pc, was clicking a bit the frontside bus. 2 active LHC WUs.
finaly, after an impressive GHz speed, the PC reboots. XP there, not disk defrag at boot. Had the same situation produced. Didn't tried again.
For 24H running, better not 2much OC.


A non qualified guess: depending the settings, the client state file is
updated every 60 sec. If exactly in this time a "unknown" cased reboot is done, the final "close file" could not be done to the file. Next time GUI starting, creating the file with the basic informations, but no more.
To put this in MY case in relationship, it's happening perhaps every 10000-20000 WU, not monitoring it.

The tonights amd is running on the same power plug as others, cl 4.09.

Not a showstopper 4me.

ID: 4595 · Report as offensive     Reply Quote
STE\/E

Send message
Joined: 2 Sep 04
Posts: 352
Credit: 1,393,150
RAC: 0
Message 4627 - Posted: 29 Oct 2004, 20:47:22 UTC
Last modified: 29 Oct 2004, 21:28:27 UTC

@ric

I got it running early this morning finally, I tried a few things but they didn't work so I was going to have to re-install the OS anyway to have HT Capability with the new CPU so I went ahead and did it. Once I did that then everything was alright, it was just a case of the newer hardware not being recognized by the OS because it was installed when the older hardware was still in the PC I guess... :)

@Michael Berger

Put a sock in it, preferably a dirty one, your growing boring and old hat, I have better things to do than carry on with the trivial banter with you. It's obvious to anyone in the Forum that you are incapable of carrying on a civil or intelligent conversation with anybody so why don't you just go cozy up with Guido, he's more your speed and style ... You both make a great pair destroying the Team Invite Threads with your Constant BS Posts in other Peoples Team Invite Threads ...

JOIN THE LHC@HOME TEAM
ID: 4627 · Report as offensive     Reply Quote
Profile Michael Berger

Send message
Joined: 2 Sep 04
Posts: 126
Credit: 49,653
RAC: 0
Message 4630 - Posted: 29 Oct 2004, 21:32:34 UTC - in response to Message 4627.  
Last modified: 21 Nov 2004, 12:17:00 UTC

ID: 4630 · Report as offensive     Reply Quote

Message boards : Number crunching : Data lost when Computer Reboot


©2024 CERN