Assignment Server Issue - GPU3 - No WS Available

[Tobit]

I'm able to connect to the Assignment Server but it is reporting that there is no work server available. I just restarted the client after many failed attempts to get work. This is a GPU3 6.40r1 client.

Code: Select all

[22:51:07] Thank you for your contribution to Folding@Home.
[22:51:07] + Number of Units Completed: 213

[22:51:07]   Successfully sent unit 00 to Collection server.
[22:51:07] + Sent 1 of 1 completed units to the server
[22:51:07] - Preparing to get new work unit...
[22:51:07] Cleaning up work directory
[22:51:07] + Attempting to get work packet
[22:51:07] Passkey found
[22:51:07] - Will indicate memory of 2046 MB
[22:51:07] Gpu type=3 species=20.
[22:51:07] - Connecting to assignment server
[22:51:07] Connecting to http://assign-GPU.stanford.edu:8080/
[22:51:08] Posted data.
[22:51:08] Initial: 0000; + No appropriate work server was available; will try again in a bit.
[22:51:08] + Couldn't get work instructions.
[22:51:08] - Attempt #1  to get work failed, and no other work to do.
Waiting before retry.

[sswilson]

Same issue here with 2 fermi cards.

[RAH]

Looks like a whole bank of servicers are down. Its called redundentcy. One stops and they all do.

I hope someone it looking into it.

HEY MODS!!

[Tobit]

It is the weekend, we might be SOL.

[stevehat1]

Same issue here with 2 fermi cards.

Ditto same here....

[sortofageek]

I have waved my little flag, realizing it is Saturday and even scientists need downtime once in awhile.

Please report back if the situation improves.

[sswilson]

I have waved my little flag, realizing it is Saturday and even scientists need downtime once in awhile.

Please report back if the situation improves.

Heh... it's not the scientists we need, it's the techs, and they're not permitted to have guaranteed weekends off.

[jay9228jay]

Yep.. same boat here, two clients waiting to send in finished work and get new work. Server 171.67.108.32 for both.

Anyone know if this is isolated to GPU3 clients or are my GPU2 clients going to be going cold soon too?

[VijayPande]

I've emailed the sysadmins to get them on this. As you have surmised, the machine is down and this is on the sysadmin side of things. Usually they are good about getting things back up quickly, but there may be some other issue here (and it is the weekend as previously mentioned).

This is pretty isolated, so it appears to have only affected GPU3, not GPU2. Since we've just introduced GPU3, there are fewer servers, so less redundancy. As we switch over, all servers will be GPU3 and there will be the full redundancy, which would also help in these situations in the future.

[sswilson]

Yep.. same boat here, two clients waiting to send in finished work and get new work. Server 171.67.108.32 for both.

Anyone know if this is isolated to GPU3 clients or are my GPU2 clients going to be going cold soon too?

My 275 pulled a GPU2 WU 15 minutes ago, and the Fermi cards have been looking for WUs well over 1/2 hr, so I suspect GPU2 servers have not been affected.

[VijayPande]

Should be coming back soon, thanks to Miles (lead sysadmin).

[Tobit]

Thanks Vijay, I just received work.

[Buck Nasty]

Just recieved work as well. Kudos to the Pande Group for a quick response to this issue on a Saturday no less!

[bollix47]

I'm again able to get work but I still have 5 result files that report "Server reports problem with unit" for 108.20 which is what the log reported prior to the server going down.

[sortofageek]

It's possible, if you still have the result files, they'll return once the servers are less busy. If not, post in the Issues with a specific WU forum giving all the relevant logs.