Startup taking too long after nettl

[Filipe Litaiff]

Hi Forumers,

We are facing a serious problem during the startup process of a N-class with 08 processors.
Normally, the whole startup takes at about 30/35 minutes, but nowadays it takes up to 03 hours.
We have checked several log files, but all of them are clear, without any kind of error message.
However, there is a message in /var/adm/syslog/syslog.log that seems to be the best clue of the problem. When the system calls the nettl process, all the following steps are delayed.
We have made a comparison with a similar server, and this process that should be started in a few seconds, takes minutes.

We are leaded to investigate the network infrastructure, from interfaces to switch, but we aren't completely sure about it.

Any help would be appreciated.

Thanks,

Filipe

[Duncan Edmonstone]

Perhaps you've got an awful lot of logging/tracing enabled...

Can you post the output of:

nettl -status

HTH

Duncan

[Eugen Cocalea]

Hi,

take a look at /var/adm/nettl.LOG00

for this, first run a:

/usr/sbin/netfmt -t 50 -f /var/adm/nettl.LOG00 >/tmp/readable_log

Post significant lines here :)

E.

[Roger Baptiste]

hi,

Turn off nettl in the startup scripts. That way you can confirm whether the problem is actually with the nettl startup; and if the problem is with nettl , you can always check after the system boots up.

nettl can be turned off by editing /etc/rc.config.d/nettl file and making the
NETTL entry 0 .

HTH
raj

[Filipe Litaiff]

Hi guys,

Thank you for the help.

I issued the command netfmt -t 50 and I had something meaningful:

***********************************STREAMS/UX*******************************@#%
Timestamp : Sun Dec 02 SAT 2001 04:07:57.966728
Process ID : 5259 Subsystem : STREAMS
User ID ( UID ) : 0 Log Class : ERROR
Device ID : 0 Path ID : 0
Connection ID : 0 Log Instance : 0
Location : 00123
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 04:07:57 401024 1 T.. 5306 9 ar_rput_dlpi: DL_UDERROR_IND, dl_dest_addr_length 8 dl_errno 4

************************100 Mb/s PCI LAN/9000 Networking********************@#%
Timestamp : Sun Dec 02 SAT 2001 04:23:52.892798
Process ID : [ICS] Subsystem : PCI100BT
User ID ( UID ) : -1 Log Class : DISASTER
Device ID : 7 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<6001> HPPCI100 10/100BASE-T driver detected bad cable connection between
the adapter in slot(Crd In#) 7 and the hub or switch.

Yeah, right, but why is the machine working properly since then? The server and the network are on production and functioning soundly. It seems that this problem, whatever it is, happens only at the startup. Does anyone have any clue? Is the lan card a good candidate for being the source of the problem?

Thanks,

Filipe.

PS: we're going to replace the card, cables and check the switch (Catalyst).