Yes got work on all 3 hosts first time, but now it's no work available on all hosts. Hope it's a demand problem and not more trouble.

I think they need to turn off task viewing again. Last time the database recovered after 3-4 hours and was able to service work requests.

I too think all the issues lately can be attributed to bringing the replica database back online.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Eric posted they had a science db crash. I do like it when they tell us what is going on. https://setiathome.berkeley.edu/forum_thread.php?id=83778#1973919

… and the recovery from this latest outage has been (and continues to be) exceptionally poor, about on par with the recovery after last weeks weekly outage.

---

Grant
Darwin NT

I'm not inclined to panic as the system hasn't been running well lately. I'm assuming that someone is at work today and already knows.

splitting isn't happening. we have 380K in RTS (good for 3 more hours)

edit: now we are down to 280K in RTS. (18:10 UTC)

. . Hi,

. . Yes I noticed that too. the GBT splitters all show as running OK but nothing is splitting, and we have seen that before ...

. . Time to kick the server. It seems we need a server bot that has one job, to kick the other servers from time to time.

. . And of course, this is with the regular outage less than 24 hours away.

Stephen

:(

Well, we're still running through the loaded tapes at much the same speed as we have been all day, but now we have a positive result creation rate of 20+, 30+ per second.

My guess (and is it only a guess) is that yesterday tapes were being split and workunits were being put into the database, but something went wrong with the housekeeping when the science database went belly up. So, to be on the safe side, they rewound the tapes and ran through them again, but most of the resulting WUs turned out to be in the database already and didn't need to be (or couldn't be) added again as a duplicate. So, no new work creation until we reached the point where it broke yesterday.

It's only here or the forum pages are extremely slow to open?

---

Nope, it is slow on the upper left coast

---

... So, to be on the safe side, they rewound the tapes and ran through them again, but most of the resulting WUs turned out to be in the database already and didn't need to be (or couldn't be) added again as a duplicate. So, no new work creation until we reached the point where it broke yesterday.

. . Right or wrong it sounds like a good hypothesis ...

Stephen

:)

. . Time to kick the server. It seems we need a server bot that has one job, to kick the other servers from time to time.

Officially , isn't that what the Transitioner for; to Transition things along?

Maybe it needs bigger boots.

---

Actually it's just pissed off that it had 2 Tuesday outages in a row on Wednesdays.

---

Actually it's just pissed off that it had 2 Tuesday outages in a row on Wednesdays.

HA ha. LOL
+1

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

UH OH getting the dreaded project has no tasks available

---

Been dealing with that for the past hour. Think the servers are getting over their daily glitch now as I am getting decent replacement work quantities.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I hope this week's shutdown will give them a chance to fix some things. The RTS keeps falling as the splitters stop, and then I'm assuming someone gives it a kick and it starts up again. So far we haven't emptied the RTS queue, but the system is still in a bad way, and I'm not sure why.

edit: I wonder if they have set the "start the splitters, refill the rts queue" at a lower point. It used to fire up when it fell to 550k or around that...and maybe they have set this number lower to give the system a longer time to do something else ??

edit: I wonder if they have set the "start the splitters, refill the rts queue" at a lower point. It used to fire up when it fell to 550k or around that...and maybe they have set this number lower to give the system a longer time to do something else ??

There's plenty of other weirdness & poor performance going on for it to just be one of the symptoms of the underlying issues, no changes to start/stop points needed.

---

Grant
Darwin NT

We are back! After 13 hours this is going to be a bumpy recovery.

Seti is coming alive again. One of my machines was able to get 113 tasks. Another machine wasn't able to get any. Hopefully the defrag of the dbase will help in the days to come.

---

~Chris

Seti is coming alive again. One of my machines was able to get 113 tasks. Another machine wasn't able to get any. Hopefully the defrag of the dbase will help in the days to come.

. . I have a feeling of deja vu ...

:)