[OpenAFS-devel] volserver tuning

[Nathan Neulinger]

Recently I started more closely monitoring our volservers for
responsiveness, especially during mass volume move and dump operations.

I've noticed that during periods where volume actions are taking place,
the volserver periodically hangs and doesn't seem to respond. Sometimes
this occurs with only a few moves taking place. 

I'm running with -p 16 on the volserver in many cases. All are on
2.4.19+ linux machines, P-III 800+, 512MB, 3ware or scsi raid.


/usr/afs/bin/fileserver -implicit rl -L \
	-udpsize 131071 -syslog -p 16 -rxpck 400 -hr 1
/usr/afs/bin/volserver -p 16 -log -syslog

All servers are on same subnet, switched 100Mbit, cisco hardware. 

Linux doesn't appear to have much useful netstat output, but it does
appear that that the "packet receive errors" number is slowly increasing
(at a tiny fraction of the rate the total)

Udp:
    46119467 packets received
    289 packets to unknown port received.
    401102 packet receive errors
    29291349 packets sent

Think the udpsize option might help that any?

rx stats: free packets 509, allocs 65133364, alloc-failures(rcv 0/0,send
0/0,ack 0)
   greedy 0, bogusReads 0 (last from host 0), noPackets 0, noBuffers 0,
selects 33321728, sendSelects 15696
   packets read: data 33626 ack 32940393 busy 0 abort 1 ackall 9182
challenge 1924 response 4399 debug 2 params 0 unused 0 unused 0 unused 0
version 0 
   other read counters: data 33066, ack 32940219, dup 231 spurious 0
dally 62
   packets sent: data 17296370 ack 13088 busy 672 abort 456 ackall 1894
challenge 4403 response 1924 debug 0 params 0 unused 0 unused 0 unused 0
version 0 
   other send counters: ack 13088, data 131200826 (not resends), resends
519994, pushed 0, acked&ignored 203527525
        (these should be small) sendFailed 0, fatalErrors 0

Only thing there that seems excessive is the number of acks. 

-----

On a related issue - a while back I remember seeing something about
support in bos for loading a new config. What's required to make use of
that? I restart each of the servers via bos restart once a week.

-- Nathan

------------------------------------------------------------
Nathan Neulinger                       EMail:  nneul@umr.edu
University of Missouri - Rolla         Phone: (573) 341-4841
Computing Services                       Fax: (573) 341-4216