topic Re: Query regarding rmail processes in Operating System - HP-UX

Query regarding rmail processes

Mei Jiao — Thu, 02 Dec 2004 22:03:33 GMT

Hi all,

From 'ps -ef | grep rmail', I can see a lot of "rmail -d root" processes with recent date/time, may I know what are these processes?

Are they normal? Why and how do they exists?

Your explanation is much appreciated, thank you very much!

MJ

Re: Query regarding rmail processes

Bill Hassell — Thu, 02 Dec 2004 23:18:03 GMT

man rmail

This is the local delivery agent for mail. Something is wrong with permissions, quite possibly with the /var/mail directory. rmail is trying to deliver messages to the root email address. It is NOT normal for these to exist for more than a second or two. Look in your email log:

tail -100 /var/adm/syslog/mail.log

This will probably be full of errors.

Re: Query regarding rmail processes

Mei Jiao — Fri, 03 Dec 2004 00:34:19 GMT

Hi Bill,

Thank you for your posting.

This is NOT normal? Ohh.. You're right the mail.log is full of this and is very huge in size:
sendmail[17231]: XBU15398: to=root, delay=1+23:10:42, xdelay=00:03:19, mailer=local, stat=Deferred

All stats = Deferred !!

Except the /var/mail directory permissions, what are the other possible causes?

Hope to hear from you again, thanks!

MJ

Re: Query regarding rmail processes

Ermin Borovac — Fri, 03 Dec 2004 01:21:01 GMT

One possible cause would be existance of /var/mail/root.lock. Some mail-related process might have created the lock and never released it. Check if anything is still using the lock with 'fuser /var/mail/root.lock'.

Re: Query regarding rmail processes

Bill Hassell — Fri, 03 Dec 2004 10:52:25 GMT

Actually, there are several reasons for a deferred mail delivery, the most likely is CPU load factor. sendmail.cf has several values to slowdown or stop mail processing when the system is under load--the reason is that mail is not nearly as important as regular processing. Here are some values to check:

# maximum number of new connections per second
O ConnectionRateThrottle=3

# minimum number of free blocks on filesystem
O MinFreeBlocks=100

# load average at which we just queue messages
O QueueLA=8

# load average at which we refuse connections
O RefuseLA=12

# maximum number of children we allow at one time
#O MaxDaemonChildren=12

# minimum time in queue before retry
#O MinQueueAge=30m

# maximum message size
#O MaxMessageSize=1000000

All of these can in some circumstances (especially heavy loads) cause delays (queued messages) in delivery.