[Info-vax] Sudden problems with slow sftp transfers and slow disk accesses

[Stephen Hoffman]

On 2014-02-18 04:42:27 +0000, gwilliams at cfa.harvard.edu said:

> Cluster of 8 Alpha boxes running V8.3 + patches, recently moved
> from one building to another.  Since the move, we've been experiencing
> odd behaviors: very slow network access (via sftp) and slow disk IO.
> ...
> MCR SCACP SHOW LAN_DEV/ALL showed numerous errors occurring over the
> past 24 hours, so this evening we replaced the switch connecting these
> 8 machines.  Errors are continuing to appear.
> 
> What am I missing?

HP suggests both ends be set for autonegotiation on for V8.3 and that's 
worth a try, but I've had autonegotiation go sideways on occasion.

If you don't have access to the switch or if it's not a managed switch, 
lock the NIC settings, and physically disconnect and reconnect the 
server to the network.

If you have managed switches, don't believe the displays, and reset the 
settings.  If that's available, I'd look to lock the settings on the 
switch and the NICs.

vLANs can do completely wacky things with connections — had one vLAN 
that was cross-connected somehow and the servers were winking onto and 
off of the network at two different sites at a pretty good clip — so 
also confirm you're all on the same vLAN.  That'll require somebody 
with access to the managed switch console to check the vLAN 
configuration.

Next step up involves confirming the cabling and the switches 
themselves are working, and that there's not a problem with a cable or 
a switchport somewhere.  I've had some Cisco switchports get tangled 
into weird states over the years, and rebooting the switch has 
sometimes cleared those.



-- 
Pure Personal Opinion | HoffmanLabs LLC