Looks like command central targets about 500K wu's ready to send, 1M wu's in progress, and 0h transitioner backlog.

What is the desired target for 'waiting for validation'?

Yes, things will change as classic transitions and if there is possibly more/better hardware. But what would be the desired target under today's configuration?

---

May this Farce be with You

Target for 'waiting for validation' should be zero. However this is unreasonable. Following any outage there will be a bottleneck in the process (ie one or more numbers or services out of whack).

This bloat will travle through the system and end at 'waiting for valicaation'. A realistic number for this process should be less than 100 I guess, and before all these latest outages (however short) it had been kept down. But now all that lag has been slumped into the validators.

Ealier this week the number was close to 500,000 and it had slowly crept back to about 20K as the validators started catching up. But the outage yesterday pushed another bubble through the system and it's back to 40K.

This will take a day or two to work through but it will return to a low number.

---

7,049 S@H Classic Credits

So the system seems to have been glitch free recently and therefore the queues should then be trending toward nominal. Instead, the validation queue is growing. Is there a reason the gradient isn't negative yet?

---

May this Farce be with You

So the system seems to have been glitch free recently and therefore the queues should then be trending toward nominal. Instead, the validation queue is growing. Is there a reason the gradient isn't negative yet?

Actually, the "Waiting for validation" value has been up and down like a yo-yo for a while now. I haven't got a clue why that is happening. :(

---

Be lucky

Neil

Yeah it seems to me that it should be going down... (It did after the latest big outage.) It did work through half a million in backlog as well as what was currently moving in.

I would think that the WFV cue would continue to shrink after outage spikes as it had done before.

Perhaps this is a result of recent upgrades. Previously the bottleneck had been that upload/downloads and/or lack of split data (work units) had kept the process of issueing/crunching and returning work slowed down enough. Now that work is being sent, processed and returned much more smoothly, the number of validators is not large enough to handle the quantity of successfully returned units.

Another thing to consider is that the SETIBOINC staff is in the process of moving all of the servers to a 1000Mbs switch to provide faster NFS access to the master data server (can't remember its name). Perhaps the validators are not yet on that switch and are communicating through the old (100Mbs or even 10Mbs-half) switch.

What we can say is that they are working. They are slowly losing ground, but they are working. The SETIBOINC staff is probably aware of this fact and may be waiting for the next big outage (planned for mid-to-late August) to move things around as the validators are not essential to getting work done, only to granting credit and preparing units to be moved to the master science database. The only concern if they can't catch up would be disk space. And by mid-to-late August we could be looking at millions of WUs in the WFV cue. (eep!)

---

7,049 S@H Classic Credits