From: Xavier Mertens (xavier@rootshell.be)
Date: Fri Jul 07 2006 - 03:39:49 EDT
After more investigations, I saw that another box (same hardware/software) had the same problem.
The memory usage grew while transferring big files via scp (>50Mbit/s) on both system. They have Intel Gigabit interfaces and use the e1000g driver. Is there knwon issues with this driver?
Regards,
Xavier
--
0 and 1. Now what could be so hard about that?
---------- Forwarded message ----------
Hi *,
I've a box running Solaris 10 6/06. 5GB of physical memory + 5GB of swap.
The box is running 3 zones and a zfs volume delegated to zones.
The server is up since 2 days and, today, 3GB of RAM were allocated in +/- 30 minutes. Then, no more big allocations.
(my SNMP monitoring and smc console reported the same values)
If I run a 'prstat -Z', I get:
ZONEID NPROC SIZE RSS MEMORY TIME CPU ZONE
2 36 657M 524M 10% 0:03:24 0.6% app5
0 59 447M 232M 4.5% 0:03:55 0.1% global
7 30 111M 69M 1.3% 0:00:41 0.0% delta5
6 37 172M 95M 1.8% 0:00:35 0.0% web5
Total: 1377M
How can I debug this issue? Memory leak? Is there new "buffers" allocated by the kernel? Help!
Regards,
Xavier
--
ESC:wq
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:40:21 EDT