Re: [perf-discuss] [networking-discuss] Irresponsiveness when lots of active network connections

2009-10-19 Thread Mika Borner
David Lutz wrote: lwp_park() was called with a timeout period. An exit with ETIME just means that the timeout expired. It isn't necessarily an error. Yeah, it seems these are just expiring connections. Looking at the problem from a different view, it looks almost like the e1000g bug 671

Re: [perf-discuss] [networking-discuss] Irresponsiveness when lots of active network connections

2009-10-19 Thread Mika Borner
Mika Borner wrote: Darren Reed wrote: Is any of this driven out of inetd? No, its a separate process, that spawns LWPs I have now trussed the irresponsive process and can see: /3:lwp_park(0xFEFD7DA8, 0)Err#62 ETIME A couple of minutes later all network interfaces

Re: [perf-discuss] [networking-discuss] Irresponsiveness when lots of active network connections

2009-10-12 Thread Mika Borner
Darren Reed wrote: Is any of this driven out of inetd? No, its a separate process, that spawns LWPs --Mika ___ perf-discuss mailing list perf-discuss@opensolaris.org

Re: [perf-discuss] [networking-discuss] Irresponsiveness when lots of active network connections

2009-10-11 Thread Mika Borner
Darren Reed wrote: What launches the loginproxy command? The loginproxy is an IMAP4 proxy. There is a setting for MaxThreads (currently set to 16384) and MaxSocketsPerThread Currently 128) The Documentation mentions for MaxThreads: "The number of available POP3/IMAP4 server threads. Each

[perf-discuss] Irresponsiveness when lots of active network connections

2009-10-10 Thread Mika Borner
Hi, on a T2 (8-core) we see irresponsiveness when having a high number of network connections, even on interfaces that do not have a high payload. Logins can take ages intrstat shows that almost all interrupts are on one strand for the payload interface (e1000g1) device | cpu5

[perf-discuss] Re: Need Some Advice Diagnosing Perfromance Problem

2006-07-04 Thread Mika Borner
> It's exactly the problem that Bart already mentioned > in a previous > reply. The sparc ide/ata driver is spinning in the > kernel in ata_wait: > > 6421427 netra x1 slagged by NFS over ZFS leading to > long spins in the ATA driver code > http://bugs.opensolaris.org/bugdatabase/view_bug.do?bu >

[perf-discuss] Re: Need Some Advice Diagnosing Perfromance Problem

2006-07-04 Thread Mika Borner
I am having the same problem on a Sun Blade 1500 with Sol10U2. I'm creating an Oracle Instance and the system performance is horrible. Any ideas? Here is the output (Sorry, attaching files did not work!): - Uptime says load average (5min): 13.09, Lockstat output (truncated): ba