System slowness, help needed on diagnosis!

[Jimbo2112]

Hi All,

I am looking for some advice on how to diagnose my system slowness. Here is the situation.

System:
E450 3x300Mhz cpu 512Mb Ram (with 50Gb external disk pack)
Sparc 20 Backup server
7 sparc 4 and sparc 5 clients
100 Mb Lan on the side of an Nt lan
Running Solaris 2.6

For the last 2 months we have been getting sporadic system speed loss. This amy only be on one terminal or it might affect all terminals. Some of the performance meters on terminals show high amounts of errors and others very little.

Can you give me advise on things to check and command line functions to get to the root of this problem?

Thanks

Jimbo

 

[Jimbo2112]

Hi,

I tried the diagnostic, but I don't really know how to read into the figures it gave me. The UDP and IP stats seem to show lots of discarded items. Is this where the error lies? See below!


TCP Statistics

481116 segments received
108 segments received in error
470775 segments received in order
2333 segments received out of order
4384 zero window probes
0 segments discarded
46518 segments sent

UDP Statistics

46914 datagrams received and delivered
28698789 discarded due to no application at the destination port
0 discarded due to other reasons
46913 datagrams sent

IP Statistics

29882431 datagrams received
2834 discarded due to IP header errors
560463 discarded due to address error
0 discarded due to an unknown or unsupported protocol
396365 datagrams transmission requests
17 discarded due to lack of resources
0 discarded due to no route to destination
0 valid routing entries discarded

ICMP Statistics

52670 messages received
0 messages received with ICMP-specific errors
3 destination unreachable
0 time exceeded
0 source quench
0 redirect
52337 echo
58406 messages sent
6069 destination unreachable
0 time exceeded
52337 echo reply

SNMP Statistics

46917 packets received
41865 GetRequests
5050 GetNexts
1 SetRequests
0 bad community names
0 operation not allowed by the SNMP community named
46911 packets sent
0 traps

Uptime : 349 days, 21:37:15
Connections Accepted: 7758
Connections Denied : 0 (not on access list)
Connections Aborted : 17
Last Host : srv_02.bma.org.uk (172.24.22.3)
Last Host TCP Port: 722
Reason : idle timeout
Time Since Abort : 1 day, 4:14:40
Connections Killed : 0
Bytes Received : 0 (current connection)
Bytes Sent : 0 (current connection)



Cheers

Jimbo
Total Bytes Received: 632678417 (since power-on or reconfig)
Total Bytes Sent : 1622 (since power-on or reconfig)

 

[hn4810]

Indeed, Printer is the problem. It make your server slow. I first guest is someone try to print from the application that the format not compatible with the printer that why you get alot of error through the TCPIP and UDP --> Created defunct --> used up the server resouce. Kill the print service --> then watch the server performance.

 

[Jimbo2112]

Hi All,

You will be pleased to know that I have fixed my system slowness. Here is what it was all about.

The printer was invovled but was actually a symptom and not a cause. The slowness was due to our printers and clients being routed through a 10/100 switch on the main network hub. This switch was sporadically slowing to the 10Mbit bandwidth when the system was under high load. I have put some clients through dedicated 100Mbit switches on the hub, and will do the rest soon.

I just have to find out now whether the slowness was due to our clients being re-routed or the network getting gradually more loaded over time.

Thanks all for your help and patience.

Jimbo

 

[mikeclark]

Remember, when using 100mb switches that it is a good idea to lock the systems to 100mb FDX and also the switch ports. Autonegotiation is dicky between sun systems and some switches.