Allocb failure on 5.0.5

[Carlitos71]

I am hoping that someone may be able to provide some insight to a problem that is happening at a dealer site that I can not fix. They are getting an error that reads "WARNING: eeE: Allocb failure during initialization" that is showing in /var/adm/messages and in the syslog and virtually filling those two file up with errors.
This system is running Openserver 5.0.5Eb with the Supplemental Release. It is a PII 450 with 128Mb of Ram. I have tried to follow the Knowledge Base to correct the problem but no luck as of yet.
So far, the RAM and NIC have been replaced. Additionally, I changed the NSTRPAGES with idtune to 2512, double what it already was. Also, I adjusted /etc/conf/pack.d/str/space.c to adjust the str_pool_size to 40 and the mblk_pool_size to 140.
But the system is still locking up in the middle of the night causing the dealer to reboot the machine in order to get back on the system.
A netstat -m is showing the failure in class 6 only.
Thanks...any and all help is appreciated.

Carl.

 

[cdlvj]

What brand NIC, and you have the latest drivers?

 

[Carlitos71]

It is an Intel Pro 100. As far as the driver, I would assume that the Supplemental would have taken care of that. Am I wrong?

Carl.

 

[cdlvj]

Not necessarily.
I would try and get the latest driver,
if that does not fix the problem, then try a 3com card.

I had a similar problem with an SMC nic, and when we replaced with a 3com, the problem went away.

 

[motoslide]

You should be able to grab the latest driver here:

ftp://ftp.sco.com/pub/drivers/OSR505/network/eeE/

Your actual problem might be caused by a device external to the SCO box. It's network adapters aren't very forgiving if they see a bunch of garbage coming in. Typical cause could be a bad hub or switch.
Is there anything which occurs at night that might be causing this problem? A big "over the network" backup process or file copy?