backup running for long period

[madhuchakkaravarthy]

Hi

server : rp 3440 ver 11iv1
for past one month backup running for long time , data backup done in msl and vsl.usually 40 gb data will get complete in 15 mins but now its taking nearly 2 hrs, since the backup running time is long in both msl and vsl i think there is no problem with library.

no nfs ,nis configured.

so what might be the problem.

regards

MC

[R.O.]

Is the backup being done through network? Maybe a half-duplex lan card could be the culprit.

Regards,

[AwadheshPandey]

use lanshow
and then lanadmin -x PPA
to see the speed.

also check with network support for slow data transfer.
you may use traceroute to library to see the connection path.

[Laurent Menase]

AwadheshPandey, lanshow is an hpux support tool which can only be run when asked by hp support.

[madhuchakkaravarthy]

Is the backup being done through network? Maybe a half-duplex lan card could be the culprit.

------------ so how can i check if it has problem or not ???

regards

MC

[madhuchakkaravarthy]

hi RO


u mean to say is

check with lanadmin

regards

MC

[Bill Hassell]

>> ------------ so how can i check if it has problem or not ???

Use the command:

lanadmin -g 0

where 0 is lan0. For other lan cards, use the lan #. Look at the stats after the Index line. For 100 Mbit or faster network cards, there should be no collisions or other errors.

[madhuchakkaravarthy]

Hi Bill

lanadmin -g 0

LAN INTERFACE STATUS DISPLAY
Thu, May 6,2010 12:41:53

PPA Number = 0
Description = lan0 HP PCI Core I/O 1000Base-T Release B.11.11.25
Type (value) = ethernet-csmacd(6)
MTU Size = 1500
Speed = 10 000 0000
Station Address = 0x13217cb4e6
Administration Status (value) = up(1)
Operation Status (value) = up(1)
Last Change = 192
Inbound Octets = 1788465033
Inbound Unicast Packets = 302519401
Inbound Non-Unicast Packets = 995443
Inbound Discards = 0
Inbound Errors = 0
Inbound Unknown Protocols = 0
Outbound Octets = 2101539065
Outbound Unicast Packets = 150059138
Outbound Non-Unicast Packets = 2729154
Outbound Discards = 0
Outbound Errors = 0
Outbound Queue Length = 2
Specific = 655367

Ethernet-like Statistics Group

Index = 1
Alignment Errors = 0
FCS Errors = 0
-------------------------------------------
Single Collision Frames = 1603492
Multiple Collision Frames = 10297943
Deferred Transmissions = 10996935
Late Collisions = 1092004
Excessive Collisions = 538
Internal MAC Transmit Errors = 467

---------------------------------------------
Carrier Sense Errors = 0
Frames Too Long = 0
Internal MAC Receive Errors = 0

so this might cause problem ???? if so how to overcome it

regards

MC

[Bill Hassell · Accepted solution]

>> Description = lan0 HP PCI Core I/O 1000Base-T

>> Single Collision Frames = 1603492
>> Multiple Collision Frames = 10297943
>> Deferred Transmissions = 10996935
>> Late Collisions = 1092004
>> Excessive Collisions = 538
>> Internal MAC Transmit Errors = 467

You have a major networking problem. This is a 1000Base-T link (called a GigE or 1Gbit LAN card) and it has not been connected to a compatible switch or the cable is too long or too old to support 1Gbit connections. The negotiated speed is 100 Mbits but is very likely HD (half-duplex) causing all the errors and a huge slowdown in performance. If you check the speed and duplex:

lanadmin -x 0

(where 0=lan0) you will probably see something like this:

Current Speed = 100 Half-Duplex Auto-Negotiation-OFF

Is your LAN switch capable of 1Gbit speed? If yes, then is the cable CAT-6? If not, then 1Gbit speed may not be possible, especially if the cable is quite long. CAT6 is also limited to less than 100 meters in length. If the switch is 1Gbit capable (and the cable is proper), then what is the port set to? It should be set to Auto-negotiate (always) for 1Gbit cards. Similarly, the 1Gbit LAN card in your system must be set to auto. This is done in the /etc/rc.config.d/ LAN files. To determine the driver for your card, use:

ioscan -kfC lan

Then edit the corresponding hp---conf file where --- is the driver name. The HP---SPEED variable should be set to AUTO_ON.