Operating System - HP-UX
1847086 Members
5202 Online
110262 Solutions
New Discussion

Re: Inetd Error : No buffer space available

 
Anil C. Sedha
Trusted Contributor

Inetd Error : No buffer space available

Hello,

I get the following error in my syslog file. Can someone pls let me know what this error is and what would be the probable resolution.

inetd[865]: ident/tcp: accept: No buffer space available

inetd[865]: ident/tcp: Died on Signal 13


Thanks
Anil
If you need to learn, now is the best opportunity
3 REPLIES 3
Brian Hackley
Honored Contributor

Re: Inetd Error : No buffer space available

Anil,

If you're on 11.11 it is possible you've hit a known issue. There is one case where it was patched up in the ARPA Transport patch PHNE_25642 or higher. In another case, a bug SR #8606247332 for which no patch is available yet. It is also possible that you can comment out the ident servicde from /etc/inetd.conf an run inetd -c to effect the change. This would disable the ident service which is is used primarly in sendmail.

Hope this helps,

-> Brian Hackley
Ask me about telecommuting!
Anil C. Sedha
Trusted Contributor

Re: Inetd Error : No buffer space available

Hi Brian,

I just got a call from HP, and i was informed that i should look whether Patches PHNE_18546 & PHNE_24419 were installed or not.

I shall however, try to see if the resolution provided by you works out or not, and maybe just go through the information on these patches.

Thanks once again for your help. If you have any info on the above listed patches, it would be a great help from you if you let me know.

Thanks & Regards,
Anil
If you need to learn, now is the best opportunity
Anil C. Sedha
Trusted Contributor

Re: Inetd Error : No buffer space available

Brian,

I got a good info for it from the HP support team.

The patch needed to be installed is PHNE_24419 which supercedes PHNE_18546

The reason is as under

PHNE_18546 introduced a design change to identd(1M), an
identification mechanism used by sendmail, that exposes
behavior which can result in severe performance problems on
any system or a TOC on an MC/ServiceGuard cluster node. This
behavior may be experienced on systems that have identd(1M)
enabled and also have a large number of concurrent TCP
connections.
- This behavior has been exhibited on systems that have more
than than 3000 concurrent TCP connections. To determine
the number of TCP connections on a system, the following
command string can be used:


Anyways, thanks for the help rendered. I shall rate your help appropriately.

Regards,
Ani
If you need to learn, now is the best opportunity