Message boards : Number crunching : Scheduler limit of 40 simultaneous trickles from a host?
Message board moderation
Author | Message |
---|---|
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
I've a host (60747) that had been disconnected from the network for just over a week. It had generated 48 trickles during that time. It was reconnected last night but kept getting no response to scheduler requests when I tried to do forced updates. I checked its stats this morning and found that 40 trickles had been recorded last night. Retried this morning and still no scheduler response, so I moved some of the trickle files to another directory and found that the scheduler only responded when I limited the number of trickles to 40. Is this a hard-coded limit? "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 5 Aug 04 Posts: 907 Credit: 299,864 RAC: 0 |
the boinc code looks like it just does a "DirScanner" on the trickle directory, so I'm not sure why the 40 cutoff seems to be there, is this a newer version i.e. 4.19? |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
Yes, it is 4.19 Carl, but I'm pretty sure the lack of response is down to a scheduler restriction. I'm certain that the sched_request_climateprediction.net.xml file sent by the BOINC client contained all 48 trickles, but don't have a copy of the file to confirm this (but I should be able to repeat the scenario towards the end of this week). "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
Seems like only 39 trickles (out of 46) were accepted this morning, with no scheduler response. All 46 trickles were in the scheduler request file. I moved 7 of the trickle files to another directory, did a project update and the scheduler responded, moved the remaining trickles back, did another update and all trickles have gone. So it definitely looks like there's something not quite right at the server end. "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
Just helped somebody who had the same problem <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=2001">here</a>. The variation in the number of trickles that are accepted suggests to me that something is timing out in the trickle handling on the server. Possibly database inserts? "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
Send message Joined: 1 Sep 04 Posts: 24 Credit: 10,865,773 RAC: 0 |
> The variation in the number of trickles that are accepted suggests to me that > something is timing out in the trickle handling on the server. Possibly > database inserts? Might well be. I've been hit by simillar problem in SETI. Boinc doesn't get scheduler response if I upload results over GPRS dial-up connection (which actually introduces latency in order of a second). It's fine if I use any other internet connection (eg. ISDN dial-up). Metod ... |
©2024 cpdn.org