1 |
james wrote: |
2 |
> On 8/21/20 4:10 PM, Grant Taylor wrote: |
3 |
>> On 8/21/20 11:01 AM, Caveman Al Toraboran wrote: |
4 |
>>> yes, i do consider re-inventing octagonal wheels. |
5 |
>> |
6 |
>> I think that it's occasionally a good thing to have a thought |
7 |
>> experiment about how $THING might be made better. |
8 |
>> |
9 |
>> It's probably good to have discussions around green feel types of |
10 |
>> replacements. |
11 |
>> |
12 |
>> But it's important to eventually assess the pros and cons of the old |
13 |
>> (as it exists), the new (as from green field), and the transition |
14 |
>> between the two. |
15 |
>> |
16 |
>> Sometimes the new doesn't warrant the transition, but it does provide |
17 |
>> an option that might be worth augmenting into the old. |
18 |
>> |
19 |
>> If nothing else, it's good to have the discussions and be able to |
20 |
>> answer why something was done or chosen to remain the same. |
21 |
>> |
22 |
>>> here, i'm just "asking" to see what makes the "safely stored" |
23 |
>>> guarantee. |
24 |
>> |
25 |
>> MTAs are supposed to be written such that they commit the message to |
26 |
>> persistent storage medium (disk) before returning an OK message to |
27 |
>> the sending server. |
28 |
>> |
29 |
>> There is some nebulous area around what that actually means.� But |
30 |
>> the idea is that the receiving server believes, in good faith, that |
31 |
>> it has committed the message to persistent storage.� Usually this |
32 |
>> involves writing the message to disk, probably via a buffered |
33 |
>> channel, and then issued system calls to ask the OS to flush the |
34 |
>> buffer to disk. |
35 |
>> |
36 |
>> Is there room for error?� Probably. |
37 |
>> |
38 |
>> Had the server made (more than) reasonable effort to commit the |
39 |
>> message to disk?� Yes. |
40 |
>> |
41 |
>> The point being, don't acknowledge receipt of the message while the |
42 |
>> message is only in the MTA's memory buffer.� Take some steps to |
43 |
>> commit it to persistent storage. |
44 |
>> |
45 |
>> That being said, there are some questionable servers / configurations |
46 |
>> that will bypass this safety step in the name of performance.� And |
47 |
>> they /do/ /loose/ /email/ as a negative side effect if (when) they do |
48 |
>> crash. This is a non-default behavior that has been explicitly chosen |
49 |
>> by the administrators to violate the SMTP specification.� Some MTAs |
50 |
>> will log a warning that they are configured to violate RFCs. |
51 |
>> |
52 |
>>> got any specific definition of what makes a storage "guaranteed"? |
53 |
>>> e.g. what kind of tests does the mail server do in order to say |
54 |
>>> "yup, i can now guarantee this is stored safely!"? |
55 |
>> |
56 |
>> It's more that they do something safe (write the message to disk) |
57 |
>> instead of risky (only store it in memory). |
58 |
>> |
59 |
>> Everything can fail at some point.� It's a matter of what and how |
60 |
>> many reasonable steps did you take to be safe.� Read: Don't cut |
61 |
>> corners and do something risky. |
62 |
>> |
63 |
>>> i guess you think that i meant that a relay should be mandatory? |
64 |
>> |
65 |
>> Sending / outbound SMTP servers /are/ a relay for any messages not |
66 |
>> destined to the local server. |
67 |
>> |
68 |
>> There is almost always at least two SMTP servers involved in any |
69 |
>> given email delivery.� All but the final receiving system is a relay. |
70 |
>> |
71 |
>>> (yes, a relay doesn't have to be used.� i'm just describing some |
72 |
>>> uses of relays that i think make sense.� (1) indicate trust |
73 |
>>> hierarchy, (2) offload mail delivery so that i can close my laptop |
74 |
>>> and let the relay have fun with the retries.� not sure there is |
75 |
>>> any other use.� anyone?) |
76 |
>> |
77 |
>> There are many uses for email relays.� A common one, and best |
78 |
>> practice, is to have an /inbound/ relay, commonly known as a backup |
79 |
>> email server. The idea being it can receive inbound messages while |
80 |
>> the primary email server is down (presumably for maintenance). |
81 |
>> |
82 |
>> Many SaaS Email Service Providers (ESPs) /are/ relay servers.� They |
83 |
>> receive email from someone and send it to someone else. |
84 |
>> |
85 |
>> A number of email hygiene appliances function as an email relay |
86 |
>> between the world and your ultimate internal email server.� |
87 |
>> Services that filter inbound email qualify here too. |
88 |
>> |
89 |
>> It is common, and I think it's best practice, to have web |
90 |
>> applications send email via localhost, which is usually a relay to a |
91 |
>> more capable hub email server which sends outbound email.� Both of |
92 |
>> these layers are relays. |
93 |
>> |
94 |
>> A relay is the same thing for email that a router is for a network. |
95 |
> |
96 |
> WOW do I love these discussions, but let me 'cut to the chase'. |
97 |
> |
98 |
> I'm proposing, via a small corp I own, to purchase up to (3) dual |
99 |
> Rasp.pi 4 setups of (2) R.Pi.4 8gig ram setups |
100 |
> and send them to the devs WE all decide on. Let's us start compiling |
101 |
> up the codes, keep it simple (for now) and implement them with |
102 |
> gentoo-users as the testers of the email services. |
103 |
> |
104 |
> |
105 |
> These discussions should be continued to everyone's benefit. However |
106 |
> there are way more than (3) folks on these threads who are most |
107 |
> capable to do this community prototyping. If WE do not act and get |
108 |
> hundreds of these deployed, email, as we know it via RFCS and other |
109 |
> standards may just disappeaar, or be relegated to the far reaches of |
110 |
> the Internet. What I have read, is standards based email services, |
111 |
> particularly by small organizations, are under extreme pressure by |
112 |
> large corporations to be marginalized out of existence. |
113 |
> |
114 |
> So any of the folks in these treads can announce publically, or send |
115 |
> me private email as to your concerns. Public is best, but, I |
116 |
> understand the needs for private communications sometimes. So yea, |
117 |
> I'll personally finaces, at least 6 months of (3) projects. |
118 |
> I'll take all input, but will make my (funding) decision, in a focus, |
119 |
> quick strategy. |
120 |
> |
121 |
> James Horton, pe |
122 |
|
123 |
I wouldn't be able to right now, just bought a new mattress set, |
124 |
mattress topper and other bed type stuff, but once I get that behind me, |
125 |
I'd be happy to buy at least one myself and compile stuff on it for |
126 |
testing. I'm not a coder by any stretch of the imagination. Heck, my |
127 |
scripts are not likely considered that by most here. I'm just not sure |
128 |
what else to call them. If I ever got bored, ran out of time or |
129 |
whatever, I could send the thing to a dev that is in the USA and easy to |
130 |
ship to, and let them play with it for a while. Only downside, my |
131 |
internet isn't dial-up but it's only a couple steps above it. Data |
132 |
transfer, especially going upstream, would be slow. Still, download, |
133 |
compile and send results shouldn't be to bad. |
134 |
|
135 |
This reminds me of that group that detects lightening. You buy this |
136 |
box, hook it to the internet and it detects lightening and sends the |
137 |
data back to their server. Then people can visit the website to see |
138 |
where the lightening is, globally at that. Thing is, getting the box |
139 |
was difficult. I wanted to buy the kit and assemble it myself since |
140 |
I've done that sort of thing before. They cost about $300 I think. |
141 |
Somewhere around there. I never did get a email that I was up on the |
142 |
list to buy one. It was pricey but at the time there was no box even |
143 |
close to me. There was sort of a hole in my area. Once hooked up, just |
144 |
keep it powered up and it requires nothing but making sure the green |
145 |
light is on every once in a while. I would like to have been able to do |
146 |
that. The data part would work even on dial-up. Data amounts were tiny. |
147 |
|
148 |
Thing about this deal, it could lead to a lot of things and benefit |
149 |
Gentoo. Who knows what someone may come up with. From what I've read, |
150 |
those little things have some get up and go to them especially for their |
151 |
size. Whether running email type software or something else, I'm sure |
152 |
any help would be accepted. Didn't someone have a guitar running Gentoo |
153 |
once?? I never did quite figure that one out. Why does a guitar need a |
154 |
computer?? :/ |
155 |
|
156 |
I suspect that there would be a few devs willing to accept the help, |
157 |
whether donating a Pi thingy or just getting testing from someone who |
158 |
has one. I can't imagine them saying no. o_O |
159 |
|
160 |
Dale |
161 |
|
162 |
:-) :-) |
163 |
|
164 |
P. S. My mattress is having issues. A couple springs are trying to |
165 |
escape. I put it off as long as I could. :-( |