FTP slow after OS upgrade to v 11.31

[Olga_1]

Hello,
We have upgraded the OS from 11.23 to 11.31.
I am not sure if this is related to the problem with very slow FTP transfer.
Any help would be appreciated.

[Tingli]

Check nic card making sure the setting hasn't been changed.

[Steven Schweda]

Which part of FTP is slow? Does it take too
long to get logged in and to get the first
"ftp>" prompt, or is looging in speedy, but
the actual file transfer is too slow?

[Olga_1]

Steven,
the actual file transfer is slow

[Olga_1]

What I noticed that after the upgrade pmtu values changed.
netstat -rnv after upgrade;
Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
127.0.0.1/255.255.255.255 127.0.0.1 UH 0 lo0 32808
xx.xxx.xx.xx/255.255.255.255 xx.xxx.xx.xx UH 0 lan900 32808
xx.xxx.x5.xx/255.255.255.255 xx.xx.x5.xx UH 0 lan2 32808
xx.xxx.xx.0/255.255.252.0 xx.xxx.xx.xx U 2 lan900 1500
xx.xxx.x4.0/255.255.252.0 xx.xxx.x5.xx U 2 lan2 1500
127.0.0.0/255.0.0.0 127.0.0.1 U 0 lo0 32808
default/0.0.0.0 10.245.48.1 UG 0 lan900 1500

before upgrade:

Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
127.0.0.1/255.255.255.255 127.0.0.1 UH 0 lo0 4136
xx.xxx.xx.xx/255.255.255.255 xx.xxx.xx.xx UH 0 lan900 4136
xx.xxx.x5.xx/255.255.255.255 xx.xxx.x5.xx UH 0 lan2 4136
xx.xxx.xx.0/255.255.252.0 xx.xxx.xx.xx U 2 lan900 1500
xx.xxx.x4.0/255.255.252.0 xx.xxx.x5.xx U 2 lan2 1500
127.0.0.0/255.0.0.0 127.0.0.1 U 0 lo0 4136
default/0.0.0.0 xx.xxx.xx.1 UG 0 lan900 1500

[Tim Nelson]

1st.. I am glad your upgrade worked..it typically does not.. or at least you start running into problems like this ;)

could you post

lanscan

then

lanadmin -x # (for each active NIC)
e.g.
# lanadmin -x 1
The link is down. The speed and other link state information
are undefined. The configuration specified for this card is:
Speed = 1000 or 100 or 10 Mbps.

[Olga_1]

Tim, Thank you for helping me.

lanscan
Hardware Station Crd Hdw Net-Interface NM MAC HP-DLPI DLPI
Path Address In# State NamePPA ID Type Support Mjr#
1/0/1/1/0/6/1 0x001560DE19A3 1 UP lan1 snap1 2 ETHER Yes 119
1/0/12/1/0/6/0 0x0016353EC126 2 UP lan2 snap2 3 ETHER Yes 119
1/0/14/1/0/6/0 0x0016353E7521 3 UP lan3 snap3 4 ETHER Yes 119
LinkAgg0 0x001560DE19A2 900 UP lan900 snap900 6 ETHER Yes 119
LinkAgg1 0x000000000000 901 DOWN lan901 snap901 7 ETHER Yes 119
LinkAgg2 0x000000000000 902 DOWN lan902 snap902 8 ETHER Yes 119
LinkAgg3 0x000000000000 903 DOWN lan903 snap903 9 ETHER Yes 119
LinkAgg4 0x000000000000 904 DOWN lan904 snap904 10 ETHER Yes 119



lanadmin -x 0
Speed = 100 Full-Duplex.
Autonegotiation = Off.

lanadmin -x 1
The link is down. The speed and other link state information
are undefined. The configuration specified for this card is:
Speed = 100 Full-Duplex.
Autonegotiation = Off.


lanadmin -x 2
Speed = 1000 Full-Duplex.
Autonegotiation = On.


lanadmin -x 3
The link is down. The speed and other link state information
are undefined. The configuration specified for this card is:
Speed = 1000 or 100 or 10 Mbps.
Autonegotiation = On.

[Tingli]

If you have the original lanadmin setup, you can reset those links back to original one.

[Tim Nelson]

post lanadmin -q

i was looking to see what your APA config is made up of.

then lanadmin -x for each nic in the lan900 agg.

e.g.
# lanscan -q
0
1
900 3 2 <<---these here
901
902
903
904

# lanadmin -x 3
Speed = 1000 Full-Duplex.
Autonegotiation = On.

# lanadmin -x 2
Speed = 1000 Full-Duplex.
Autonegotiation = On.

[Olga_1]

Tim,
APA was set up to have lan0 and lan1 in lan900

but now lanscan -q shows
1
2
3
900 0
901
902
903
904

and
Parent Link Aggregate -1
Speed 100 Mbps Full Duplex
MTU 1500
Virtual Maximum Transmission Unit 0
MAC Address 0x001560de19a2
Features VLAN,CR4,CS4
Network Management ID 6
Load Distribution Algorithm HOT_STANDBY
Admin Status UP
Link Aggregate Health 50 %
List of ports in Aggregate 0,1
List of active ports 0
List of not ready ports 1

[Tim Nelson · Accepted solution]

well then your Aggregate is not working.

If you were depending on the 200Mb/s then right now you are running 1/2..

I would review all your NIC connections and be sure that you have the right ones, maybe the numbering is off.

As a note, I see a couple 1000Mb NICs, maybe time to use them if your backbone supports them ?

I cannot be sure if this is completely slowing down your FTP, APA does not fully balance the work anyway. and your config says it is in HOT-STANDBY, albeit with a missing NIC.

I would also check on all the NIC firmware.. maybe they became buggy with 11.31 and need to be updated as well?

Have not used 100Mb nics in awhile so I am not familiar with 11.31 issues.

[Steven E. Protter]

Shalom,

Yet another reason to cold install, not upgrade.

Ideas:
1) Check the system has a line for self resolution in /etc/hosts and /etc/nsswitch.conf actually resolves files first.
2) Check for a recent bi-annual patch update to 11.31, if available, get it installed.

SEP

[James R. Ferguson]

Hi:

> Yet another reason to cold install, not upgrade.

That may have been very true in the days of the 10.20-to-11.0 transition, but the upgrade process has improved greatly. Search the forum for some success stories.

> Ideas: 1) Check the system has a line for self resolution in /etc/hosts and /etc/nsswitch.conf actually resolves files first.

In a cold-install there is *no* '/etc/nsswitch.conf' file deposited. You need to copy one of the templates like '/etc/nsswitch.files' and usually modify it to your needs. Seems like your more likely to miss this in a cold-install :-)

> 2) Check for a recent bi-annual patch update to 11.31, if available, get it installed.

Statements like this are throwing stuff at a wall to see what sticks. A cold-install or an update will apply the standard patch bundles in the OE (Operating Environment) media. While certainly their may be additional patches to correct last minute problems, statements like this remind me of those who blame operating systems for all of their application code and/or local configuration problems. A more focused approach would be to search the Patch Database by keyword and examine the corresponding patch notes.

Regards!

...JRF...

[rick jones]

The specifics/syntax of the attached is more Linux-specific, but the ideas should translate to HP-UX - I tend to start by looking for lost packets and then figuring-out why those packets were lost.

BTW, this FTP transfer, is it a local LAN transfer or is it a WAN transfer?

[Viktor Balogh]

Hi Olga,

attached is a script which prints the network usage every second. Maybe it will be useful for network testing.

[Olga_1]

Thank you very much for all the replies.
Patches are up to date on the server.
We did loose one NIC card (lan1) and HP tech is coming on site today to replace it.
That however does not explain slow FTP. As we did not have a load balancing set up. Our APA is configured: HP_APA_HOT_STANDBY[0]=on.

FTP speed on this server:
229376 bytes received in 5.54 seconds (40.42 Kbytes/s).

On the rest of the servers that are on the same network

63607842 bytes received in 1.36 seconds (45706.96 Kbytes/s)
Any ideas?

[rick jones]

link-layer statistics (lanadmin -g mibstats ) for the interface(s) in use might be helpful.

[Olga_1]

Nic card replaced, however problem with FTP is most likely connected to swapping. We are running out of memory.