amd64/145654: amd64-curent memory leak in kernel

Andrey Smagin samspeed at mail.ru
Mon Apr 12 15:00:05 UTC 2010


>Number:         145654
>Category:       amd64
>Synopsis:       amd64-curent memory leak in kernel
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-amd64
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Mon Apr 12 15:00:04 UTC 2010
>Closed-Date:
>Last-Modified:
>Originator:     Andrey Smagin
>Release:        amd64-current 2010 mart xx
>Organization:
DiP Interactive
>Environment:
FreeBSD rizhskaya8.vvt 9.0-CURRENT FreeBSD 9.0-CURRENT #2: Mon Mar 29 23:45:02 UTC 2010     root at rizhskaya8.vvt:/usr/obj/usr/src/sys/SAM  amd64
>Description:
I switched i386->amd64 on my server, with full new ports tree, 
save only configs. Under heavy disk IO, "Wired" memory increased and 
kernel panic in malloc inside IPFW. Uptime back proportional of disk IO load,
with transfer large files from one HDD to another,paniced from 1 to 3-4 hours. With no IO uptime about 2-5 days. It may be ZFS related problem because, 
I switch /usr and /var to ZFS. I try to fix problem by configuring ZFS in 
loader.conf, but nothing help.
loader.conf:
vfs.zfs.arc_max="512M"
vfs.zfs.arc_min="256M"
vfs.zfs.vdev.cache.size="128M"
panic always in:
panic()
kmemmalloc()
uma_large_malloc()
malloc()
ipfw_ctl() at ipfw_ctl+0xb4. 
rip_ctl_output() at rip_ctl_output+0x9f
...
>How-To-Repeat:
start copying large size of data from UFS to ZFS or vice versa. "Wired" memory  increased value and after some hours value above 2GB then hang in malloc always in IPFW.

>Fix:


>Release-Note:
>Audit-Trail:
>Unformatted:


More information about the freebsd-amd64 mailing list