Re: [exim] Failed to get write lock on smtp?

Top Page
Delete this message
Reply to this message
Author: Philip Hazel
Date:  
To: Marten Lehmann
CC: exim users
Subject: Re: [exim] Failed to get write lock on smtp?
On Thu, 19 Oct 2006, Marten Lehmann wrote:

> I have a smtp-transport using lmtp. In the mainglog I got several like this:
>
> Failed to get write lock for /var/spool/db/wait-mailbox_lmtp.lockfile:
> timed out
> What is this write lock for? mailbox_lmtp is the name of the transport.


For some reason the LMTP delivery has suffered a temporary error. Exim
is trying to update its database to record this fact so that it knows
when to try again. The name "wait-mailbox_lmtp" indicates that it is
data for addresses that were deferred by the mailbox_lmtp transport.

The fact that Exim cannot access its hints database is worrying.

> Is this transport treated as a remote delivery so remote_max_parallel
> applies?


Yes. It *is* a remote transport; there is no reason why it shouldn't run
deliveries to multiple hosts in parallel.

If you are still having this problem, try running a delivery with
debugging turned on to see if that helps:

exim -d -M <the message id>

The output goes to the standard error - you may want to send it to a
file. The other thing that could be checked is whether some other
process is holding a write lock on the file, and if so, what it is
doing.


-- 
Philip Hazel            University of Cambridge Computing Service
Get the Exim 4 book:    http://www.uit.co.uk/exim-book