On 2013-02-02 18:26, Jake Young wrote:
Oddly enough I never had any issues with the automounter or nfs share I was
backing up the V445 to. One issue was with the ssh connection between my
OI desktop and the V445 dropping out (note there is very little text output
in my backup script) after an hour or two. My other issue was having the
Speculating on these points, I'd suggest that:
1) NFS is by default over UDP, which may be more resilient to dropped
packets and adaptation to smaller sizes can be natively bundled.
The higher-layer programs above UDP (i.e. NFS) have to detect and
retry the transmissions, however - unlike TCP which should handle
this sort of problems but may lag until it delivers its payload of
bytes (the "sliding window", usually several typical packets in
size).
2) SSH drop-out may be timeout-related... See if your server and/or
client can use (TCP)KeepAlive, or if the script could output more
data to keep the connection "active". Timeouts might be related
to the IPFilter (NAT entry age, size of session bucket, etc).
3) Speaking of the latter, we had a problem on our firewall with the
default IPFilter session bucket being very small. I am not sure
if this was solved in later releases, we applied a fix like this
at the time (and it's still running):
[root@newgw /lib/svc/method]# gdiff -bu ipfilter.snv_129-orig ipfilter
--- ipfilter.snv_129-orig 2009-11-27 01:07:18.000000000 +0300
+++ ipfilter 2010-06-29 14:07:36.034883267 +0400
@@ -153,7 +153,12 @@
create_services_rules || exit $SMF_EXIT_ERR_CONFIG
[ ! -f ${IPFILCONF} -a ! -f ${IPNATCONF} ] && exit 0
- ipf -E
+
+ ### Enforce and display state-table sizing
+ ### Jim Klimov, 2009-2010
+ ipf -D -T
fr_statemax=72901,fr_statesize=104147,fr_statemax,fr_statesize -E -T
fr_statemax,fr_statesize
+ # ipf -E
+
load_ippool || exit $SMF_EXIT_ERR_CONFIG
load_ipf || exit $SMF_EXIT_ERR_CONFIG
load_ipnat || exit $SMF_EXIT_ERR_CONFIG
In fact, in OI 151a7 I see that the values are set to 50000, so
it seems, while on older Sol10u8 and SXCE it is around 4000.
Our values were from trial-and-error somewhat; and I think there
was some math to them as well (google or ask Darren) ;)
HTH,
//Jim
_______________________________________________
OpenIndiana-discuss mailing list
OpenIndiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss
