On 10/7/2011 3:41 AM, Bernhard Schmidt wrote: > Basically the only problem with postfix here is that I cannot have > queue_minfree > 2GB to be on the safe side, so I don't know how to avoid > this problem.
There is a simple solution here, Comp Sci 101 type stuff, which Wietse has mentioned many times on this list: If your input rate exceeds your output rate by such a huge margin that the resulting spooling of 2.5m emails breaks your filesystem, simply decrease the input rate into Postfix. You mentioned in your first post that this is a bulk mailing scenario. I dare say that even if you are able to get 2.5m mails spooled on a single host, with Postfix' default smtp settings, your IP will get blacklisted pretty darn quickly at many receiver sites due to the deluge. So now you have another problem, which can't be fixed by switching/tweaking filesystems. The only sane thing to do here is to reduce the input rate or parallelize the workload across multiple hosts. If you must move that many emails in a short period of time, you need to build an outbound farm and distribute the emails evenly across many Postfix instances on many hosts. This is what everyone else does for this type of workload. Search the list archives for specifics. Such setups have been discussed in detail here. If I may make a purely subjective comment: 2.5m spooled emails on a single host is insane. -- Stan
