[Tfug] Need help on a server performance issue...

[Jim March]

Arright, some updates after I visited today.

Every test I've thrown at the disks says they're fine.  A little slow
but about what I'd expect on craptastic hardware.

The two "hubs" are indeed switches - consumer-grade 10/100 but that
should work.  It's possible one of 'em is fubar of
course...but...hmmm...

ifconfig -a says:

---
[crappybox]@fileserver:~$ ifconfig -a
eth0      Link encap:Ethernet  HWaddr 00:10:dc:72:68:d6
          inet addr:192.168.168.250  Bcast:192.168.168.255  Mask:255.255.255.0
          inet6 addr: fe80::210:dcff:fe72:68d6/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2132681 errors:0 dropped:0 overruns:0 frame:0
          TX packets:2231204 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1040247421 (1.0 GB)  TX bytes:1641024208 (1.6 GB)
          Interrupt:23 Base address:0xc800

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:449 errors:0 dropped:0 overruns:0 frame:0
          TX packets:449 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:42328 (42.3 KB)  TX bytes:42328 (42.3 KB)

[crappybox]@fileserver:~$
---

Looks OK to me.

I ran IPTraf for a while.  Once in a while it'll throw out a red line
below on what I assume is a collision, but it's not happening enough
to be worrisome I don't think:

---

UDP (147 bytes) from 192.168.168.121:17500 to 192.168.168.255:17500 on eth0
UDP (69 bytes) from 192.168.168.75:5353 to 224.0.0.251:5353 on eth0
UDP (198 bytes) from 192.168.168.250:631 to 192.168.168.255:631 on eth0
UDP (147 bytes) from 192.168.168.121:17500 to 255.255.255.255:17500 on eth0
UDP (147 bytes) from 192.168.168.121:17500 to 192.168.168.255:17500 on eth0

But maybe I'm reading it wrong?

I guess I'll have to wait for everybody to go home and start shutting
off boxes one by one until the server-to-client copy speeds jump up?!?
 But if I'm reading IPTraf right, that's not the issue either...?

I'm kinda stumped...

Jim