What to do when Status is WAITING

[twizforpros]

Hello:

After running for a week or more without incident, twice now I've waked up my computer after it's worked overnight to find a status of WAITING with a progress of 100%. I can see multiple "failures to connect" in the log, both WS and Collection servers, plus "unable to get assignment" messages. Based on another forum entry where a user "refreshed" his slots to recover, I deleted and re-added my two slots (CPU and GPU) and things began working again. BUT..., should I have left things alone? Would things have worked out on their own? Do I risk losing work by deleting/re-adding slots? I did look at the server status page but the referenced servers seemed to be ok.

FAH 7.3.6
i7-4770k 3.5 OC'd to 4.4
16GB
EVGA (nvidia) GTX760 w/4gb

Thanks

[Jim Saunders]

All of mine get a new unit automatically after the work assignment server comes back up. Chances are if you're not getting WUs then most likely no-one else is either.

Jim

[P5-133XL]

When there are connection difficulties, whatever the cause, the client automatically tries to reconnect but each time it does, it will wait longer till the next retry. If it lasts long enough with enough retries the wait can get to be a long time. If you look at the advanced control there is a next attempt field in the selected work unit area that will list the time till the next retry. You can manually reset the time between retries by pausing and then unpausing each stalled slot or simply restart the client.

If you delete the slots then the client will recognize that there is no appropriate slot for that data and clean it out of existence after you save the new config and leave. If you delete the slots and recreate them in the same config action then when you save and leave then nothing will have changed and the data is safe. In both cases, the act of saving and leaving config will be a signal for the client to restart which will reset the timers.

P.S. I was corrected. Unless you are experiencing bug #983: "No error recovery following an unknown Communications Error; client needs to reset communications".

[7im]

Quit and restart the client (task bar icon) resets the clock, with no chance of data loss.

[twizforpros]

When there are connection difficulties, whatever the cause, the client automatically tries to reconnect but each time it does, it will wait longer till the next retry. If it lasts long enough with enough retries the wait can get to be a long time. If you look at the advanced control there is a next attempt field in the selected work unit area that will list the time till the next retry. You can manually reset the time between retries by pausing and then unpausing each stalled slot or simply restart the client.

If you delete the slots then the client will recognize that there is no appropriate slot for that data and clean it out of existence after you save the new config and leave. If you delete the slots and recreate them in the same config action then when you save and leave then nothing will have changed and the data is safe. In both cases, the act of saving and leaving config will be a signal for the client to restart which will reset the timers.

P.S. I was corrected. Unless you are experiencing bug #983: "No error recovery following an unknown Communications Error; client needs to reset communications".

Thanks for this. Sorry to be so dense but, how does one "pause" and "un-pause" a slot? Whoa! Hang on, I moved the slider bar all the way to the left and I see the status changes to "Pause" Answered my own question. Thanks again.

[twizforpros]

Quit and restart the client (task bar icon) resets the clock, with no chance of data loss.

Thank you for this. The Client is very very smart if he can be shut down and re-started without missing a step.

[bruce]

That's not "dense" since it's obscure. [I'm assuming Windows but an answer for other platforms can be provided.]

The client restart is a better method, but I'll answer the question you asked anyway. You might need it someday.

Right-click on the systray Icon and select Advanced Control.
Under "Folding Slots" right-click on each slot and choose "Pause"
Once everything shows "Paused" right click again and choose "Fold"

[bruce]

The Client is very very smart if he can be shut down and re-started without missing a step.

In truth, the client will miss a few SMALL steps but nothing major (unless you restart before it's finishing shutting down).

[twizforpros]

Excellent. Thanks. I'll go with the restart as it seems to be the least risky. Oops - looks like I need to give it time to shut down gracefully before I restart. Five minutes enough?

[Jim Saunders]

Excellent. Thanks. I'll go with the restart as it seems to be the least risky. Oops - looks like I need to give it time to shut down gracefully before I restart. Five minutes enough?

Plenty. I've watched it pause (using 7.2.9 at least) and I can't recall it taking more than a minute to go from a pause command to indicating that it's paused.

Jim

[bruce]

One minute is often enough but it can be borderline, depending on factors you don't need to understand. Two minutes is enough. Longer can't hurt but is unnecessary.

[Jim Saunders]

One minute is often enough but it can be borderline, depending on factors you don't need to understand. Two minutes is enough. Longer can't hurt but is unnecessary.

Learn something every day.

Jim

[twizforpros]

Thanks, everybody. Extremely helpful!