Heads up: WCG update can error out your machines entire queue!

**fallwind** · 03-01-2011, 02:28 AM

knreed explaining the update: https://secure.worldcommunitygrid.or...d_thread,30943

My post showing what happened to my machines: https://secure.worldcommunitygrid.or...d_thread,30946

Sekerob posed a possible cause in post# 5. THere's no need to take action at this point, if you machine hasn't errored out yet then it's likely fine. If it has errored out, let it crunch and return some results and your cache will refill on it's own.

**Jaco** · 03-01-2011, 02:33 AM

Haven't noticed anything unusual here , but thanks for the heads up

EDIT : My 980X has the same issue :-(

<core_client_version>6.4.5</core_client_version>
<![CDATA[
<message>
app_version download error: couldn't get input files:
<file_xfer_error>
<file_name>wcg_hcmd2_maxdo_6.15_windows_intelx86 </file_name>
<error_code>-120</error_code>
<error_message>signature verification failed</error_message>
</file_xfer_error>

</message>
]]>

**haschioz** · 03-01-2011, 05:32 AM

i had a strange thing happen to me two days ago. i crunch on my lappie and when i got the lappie was idling. wcg told me i reached the quota and all my wu's errored (10 or so) and went on for like 20 hours.

**jcool** · 03-01-2011, 05:56 AM

Epic fail across the board here, all queues effed up.

Thanks for posting though, I thought at first that maybe something was wrong with my SR-2

**64dragon** · 03-01-2011, 05:54 PM

dang, which I had checked here sooner. checked my email this morning before going to work (don't normally) and saw that my pc was idle and all WU's errored out, figued my overclock became unstable or something so I haven't been crunching all day

at least I only lost 50+ WU's. At least I didn't loose the rest of the night running Prime95.

**D_A** · 03-01-2011, 06:21 PM

Not a blip here. George just luck I guess.

**nanoprobe** · 03-01-2011, 06:27 PM

Originally Posted by D_A

Not a blip here. George just luck I guess.

None here either. *Big sigh of relief* Wonder why only some were hit?

**D_A** · 03-01-2011, 06:35 PM

Something to do with how the security certificates are handled at a guess, but that's about as far as I can work it out.

**STEvil** · 03-01-2011, 06:55 PM

got hit with that last night, it even managed to crash utorrent when it did it somehow.

**Biker** · 03-01-2011, 07:09 PM

Yup it got me as well and trashed 2 machine's queues :/

I now have over 60 pages pending as a result and my 2 most productive machines are no longer trusted to process DDDT2 WU's (no WU's available for this project)

**retro77** · 03-01-2011, 09:58 PM

What project are the errors occurring on?

**STEvil** · 03-01-2011, 10:19 PM

everything.

its a security key bug. Notice the "6.40" at the end of the "appication" name? Thats the new key version, everything that crashed was 6.12 or something.

**josh1980** · 03-01-2011, 11:31 PM

How come there are so many problems with reliability of the WCG network? It seems like regularly we hear about a failed file server, corrupt file allocation table, etc.

**stoneageman** · 03-02-2011, 01:04 AM

Wonder if they started 'testing' this on the 25th, as that's when I started having multiple hcc errors as described with all my linux rigs

Thought it could be a virus

. Seems to have sorted itself out now

**D_A** · 03-02-2011, 03:28 AM

Originally Posted by josh1980

How come there are so many problems with reliability of the WCG network? It seems like regularly we hear about a failed file server, corrupt file allocation table, etc.

Well given that there are several hundred thousand users requesting and returning work daily I'd think they don't do too bad. Yes they might be able to do better in absolute terms, but that's a lot of database transactions and sooner or later you're going to get a bad something no matter what you do.

**fallwind** · 03-02-2011, 03:31 AM

4 out of 6 of my farm rigs blew their caches out. It appears to be random as all 6 are identical software wise and very similar in hardware. Type of WU's didn't matter either as I lost queues of DDDT2, CEP2 and HFCC alike. Didn't lose much crunch time and the rigs are back on the reliable list and receiving repair WU's so nothing to get too upset about here.

**Nikolasz** · 03-02-2011, 04:16 AM

Agree, dont press update!

**STEvil** · 03-02-2011, 04:41 PM

Originally Posted by Ego

Agree, dont press update!

The fix is to press update to manually flush the cache..

**D_A** · 03-02-2011, 06:31 PM

That's kind of the point. If you DON'T hit "update" the it will still happen just at some time that will probably be even less convenient. Better to get it over and done with when you're there to get it sorted.

**Tom128** · 03-03-2011, 01:15 PM

I can't tell if I am getting hit by this or if I have an unstable overclock. I have 30 WU's that have just errored on a machine I just put up. It's first two days worth of units were valid, and now all of the new ones are showing up as errored.

The error:
autogrid4: WARNING: I just prevented an attempt to take the arccosine of 1, a value greater than 1.

That sounds COMPLETELY different from what you guys are talking about, and seems like a stability error. HOWEVER I am looking at the quorum for the WU's and there seems to be alot of other people getting errors on the same units.

I am at work right now so i can't take that machine down, but I am going to take it out and do stability tests. Very sad about 30 errored WU's, it seemed completely stable before.

EDIT:
Every single WU that I errored out had another person error on, unless I was the only with finished. Also, about half of my errors look like they errored out instantly with an out of memory error. I have 4Gig of memory on the machine so that is a surprise to me. Do not like.

**haschioz** · 03-03-2011, 01:55 PM

only thing i would try is check temps downclock to stock and see what happens. if it persists it likely going to be the wu, if not oc as far as its stable.

**Tom128** · 03-04-2011, 06:25 AM

Well my system was unstable. Apparently very unstable. I've used a program called K10Stat to clock my Phenom II's from windows for a few years now, as it's much faster than using the BIOS. Long story short it does not seem to like my X6's, perhaps just on that motherboard and/or it's BIOS, but the voltages it was apply and reporting were pretty far off, around 0.05v from the settings I was using. Very confused because it was passing prime95 for a few hours, and it did 35K+ WCG points without issue and THEN completely bombed out.

I stopped using it and went back to good old fashion BIOS clocking and have it running so far at 3.5Ghz @ 1.25v. I'm going to make sure it's ultra super deluxe stable before I put it back in.

Now the question is what kind of impact the ~15 errored and ~15 insta-canceled WU's is going to have on what WCG will send the machine once I put it back in.

**David_L6** · 03-05-2011, 02:52 AM

Originally Posted by Tom128

Now the question is what kind of impact the ~15 errored and ~15 insta-canceled WU's is going to have on what WCG will send the machine once I put it back in.

Not much, if any.

**Johnmark** · 03-05-2011, 10:08 AM

Originally Posted by Ego

Agree, dont press update!

Had the same problem on two machines, the only way I could get the client/server to start talking again was to do an update. Working fine since then?

**Tom128** · 03-05-2011, 11:57 AM

Originally Posted by David_L6

Not much, if any.

Seems correct, for about an hour the machine got zero work units, and then the flood gates opened and I got several dozen.

Thread: Heads up: WCG update can error out your machines entire queue!

Thread Tools

Search Thread

Rate This Thread

Display

Heads up: WCG update can error out your machines entire queue!

Bookmarks

Bookmarks

Posting Permissions