Issues with the Client and Core

[Spazturtle]

1) bigadv server goes down for updates and all bigadv rigs go idle, why not assign them normal WUs? or at least say in the logs that the server is down for upgrades.

2) folding core are broken on linux with etx3 and etx4 (more etx4), they make an un-necessary number of file-system calls causing the wu to sit for up to an hour (in the case of bigadvs, minutes for normal WUs) before being uploaded.

[Jesse_V]

There are several threads about ext4, for example viewtopic.php?p=228279 and viewtopic.php?f=55&t=17972

[Spazturtle]

There are several threads about ext4, for example viewtopic.php?p=228279 and viewtopic.php?f=55&t=17972

Its not just ext4 though, it happens on any journaled file system that uses correct semantics, its just harder to see it happening on etx3.
One of the main problems with this is that it actually damages some peoples hardware, if you fold on an ssd it can causes weeks worth of wear in a few days.

[Joe_H]

1) bigadv server goes down for updates and all bigadv rigs go idle, why not assign them normal WUs? or at least say in the logs that the server is down for upgrades.

Usually there is other SMP work that bigadv systems can process, if you have specific examples of this not happening please post them. This would include system and configuration information and logs showing which servers were contacted for work assignments. There may be other issues involved in why a particular setup did not get assigned regular SMP, but in general they are supposed to get assigned SMP if bigadv work is unavailable.

[Spazturtle]

Usually there is other SMP work that bigadv systems can process, if you have specific examples of this not happening please post them. This would include system and configuration information and logs showing which servers were contacted for work assignments. There may be other issues involved in why a particular setup did not get assigned regular SMP, but in general they are supposed to get assigned SMP if bigadv work is unavailable.

I don't have access to the logs on my machine at the moment but the error was "time:ERROR: Exception: Server did not assign work unit"
I was on 128.143.231.201 IIRC.
Me and two other people in my team with bigadvs rigs had this issue, it would just keep trying the same server, I don't know what server they were on. I use 7.3.6 and one of the others uses v6.

[Joe_H]

We would need enough of the log to also see which AS was connected to before the system was assigned to 128.143.231.201. There was a similar problem recently where a number of folders on normal or advanced SMP were getting assigned to a WS with only beta WU's available and not moving on to a server with suitable work. In that case an incorrect setting on the WS was corrected and fixed the issue. So if you get stuck this way for more than say 20-30 minutes, a report may help track down an issue on either an AS or WS.

[Spazturtle]

Hi I have uploaded my log to pastebin, it starts at the point of finishing the WU (Next unit percentage is set to 100) and ends when I change config to get normal units.

http://pastebin.com/egkxA9km

[PantherX]

Please note that if it was 128.143.231.201, then there were some issues with it which were resolved -> viewtopic.php?f=18&t=23898