30 Sep
2022
30 Sep
'22
8:13 p.m.
The recent question reminds me - since leaving for another VPS provider,
the various VPSes have had various difficulties with getting emails from
them accepted by the dominant Google / Microsoft / Yahoo services.
In one case, the content - requested activity alerts from a message board -
hasn't changed, and the volume is way *way* under anything suspiciously
spambot-y - a hundred emails a week, not per second - but Microsoft have
simply refused to accept email to users that have Hotmail / Outlook
accounts, despite the only significant change being the IP address it comes
from.
This is even after jumping through all of the hoops in the sometimes very
outdated Microsoft 'help' (sic) pages: I strongly suspect that they are
blocking the new provider's whole IP address space (or nearly so...)
Despite having various filters set up to say 'no, it's not spam, how many
times do I have to tell you', Gmail still keeps wanting to label fail2ban,
logwatch, and other system report emails as spam.
I look at
https://cfenollosa.com/blog/after-self-hosting-my-email-for-twenty-three-ye…
and go 'at least it isn't just me' and 'I hope someone takes the
oligarchs
to someone with teeth', but before I lose all my hair..
.. can you still get emails delivered from Bitfolk VPSes to the three
reliably to inboxes?
Ian
30 Sep
30 Sep
8:32 p.m.
Hi,
On Fri, Sep 30, 2022 at 10:13:56PM +0200, Ian via BitFolk Users wrote:
the various VPSes have had various difficulties with
getting emails from
them accepted by the dominant Google / Microsoft / Yahoo services.
All of the large email service providers have tightened up their
policies in the last 12 months, but probably none more so than
gmail who now require at least an SPF and/or DKIM pass.
.. can you still get emails delivered from Bitfolk
VPSes to the three
reliably to inboxes?
There's 328 addresses on this list, 98 of them are at g(oogle)?mail;
the only bounces received today were from force9.co.uk.
We still have problems from time to time.
If you still have a bitfolk VM, you could try redirecting some or
all ails through there to see if you do better (don't forget to
update your SPF!). You must absolutely make sure that you have your
reverse DNS, SPF and DKIM all squared away first though. And don't
forget both IPv6 as well as legacy IPv4.
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
8:38 p.m.
On Fri, Sep 30, 2022 at 08:32:48PM +0000, Andy Smith via BitFolk Users wrote:
You must absolutely make sure that you have your
reverse DNS, SPF
and DKIM all squared away first though. And don't forget both IPv6
as well as legacy IPv4.
mailop is a mailing list for email admins that can give more
extensive advice, though it is really for much bigger senders than
your average hobbyist (or all of BitFolk):
https://www.mailop.org/
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
11 Oct
11 Oct
10:21 p.m.
On 30/09/2022 21:38, Andy Smith via BitFolk Users wrote:
mailop is a mailing list for email admins that can
give more
extensive advice, though it is really for much bigger senders than
your average hobbyist (or all of BitFolk):
https://www.mailop.org/
Five years ago I had cause to ask questions on that list and
it *was* helpful and reviewing my mail archive I see that I
also had some fairly in-depth *off-list* replies.
In my case one of my customers is on hetzner.de: reason
being that he's got 16TB of photos of motorbikes and I
cannot find anywhere else that can provide that amount
of storage without it costing an arm and a leg (and
coming with vast amounts of compute power we don't need). [1]
The problem was that the IP address was blackholed as
being in a dodgy block. I thought that this was like
being issued an ASBO because your someone else on the
street threw noisy parties.
But overall the mailops attitude was "your problem,
not ours".
It's annoying - a btinternet customer has a reasonable
expectation that a *desired* email - order confirmation -
should be delivered.
[1] In particular one can rent a second-hand server that's
already got 4 x 6TB disks (which I mount as RAID, obvs.)
for much less than the cost of provisioning one from new.
I guess we are still happy for now but suggestions for
alternatives are welcome. Andy - feel free to quote!
12 Oct
12 Oct
11:20 a.m.
Hello,
On Tue, Oct 11, 2022 at 11:21:16PM +0100, Dom Latter via BitFolk Users wrote:
The problem was that the IP address was blackholed as
being in a dodgy block. I thought that this was like
being issued an ASBO because your someone else on the
street threw noisy parties.
But overall the mailops attitude was "your problem,
not ours".
I think that was their attitude but not necessarily because they
agree it should be like that, but maybe because after long
experience they realise it IS like that and it's outside their power
to change it, they're just telling you how it is.
The users of the largest inbox providers aren't really paying for
the service. Apart from the personal information they give up for
marketing purposes. There's little margin for going the extra mile
to separate spam from ham, big hammers are used, and still the users
stick with the service and blame senders for important mails that
don't arrive.
Looking at the issue from another angle, why is it that people who
are so bothered about being able to receive good mail from bad
networks pay little or nothing for the service that delivers this
supposedly important email? Rhetorical question - it's human nature.
It's annoying - a btinternet customer has a
reasonable
expectation that a *desired* email - order confirmation -
should be delivered.
Yet they trust their email to the lowest bidder. BT outsourced their
email to one of the big players to reduce costs. People expect a
barely functional email address with their Internet connectivity and
that is all they get.
I too have some networks blocked in my mail servers because there's
just too much spam from them. Mail from those networks won't be
scored based on content, there won't be any opportunity for a DKIM
match to allowlist the domain, the connection will just get dropped.
The difference is that a legit sender could still contact
postmaster(a)bitfolk.com and I'd work out a way to allow the email.
None of the big email providers have any customer service to do
that.
While people prefer to receive their email at places that choose to
not provide customer service, they have to appreciate that this
doesn't allow any flexibility in niceties like allowing legit email
from bad networks. If they don't like that they should self-host or
go with a provider that is more flexible. But they won't because
that usually costs time and/or money.
For most people email is already dead and is only used to sign up
for accounts on web sites. Eventually maybe even that use case will
be solved and then email is truly over. I say that as someone who
has had an email address since 1992, and still uses the one I've had
since 1998.
What advice was mailop supposed to give you when no one who read
your email was in a position to force HotGoogle to hire hundreds of
people to work as postmasters, remove all their IP-level blocks of
OVHetzner, spend their working weeks adding allowlists for good
senders at those places and start charging £100/year to every free
account to pay for it? When neither the shareholders nor customers
of HotGoogle want that.
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
14 Oct
14 Oct
2:17 a.m.
On 11 Oct 2022, at 15:21, Dom Latter via BitFolk Users wrote:
[1] In particular one can rent a second-hand server
that's
already got 4 x 6TB disks (which I mount as RAID, obvs.)
If you’re squeezing 16TB of storage out of 24TB raw, it implies the use of single parity
RAID. I hope you’ve got good backups :)
Ant
3:10 p.m.
On 14/10/2022 03:17, Anthony Newman via BitFolk Users wrote:
On 11 Oct 2022, at 15:21, Dom Latter via BitFolk Users
wrote:
It does. Backups are the customer's responsibility :-}
I've not considered single parity raid to be risky - what do other
people think?
[1] In particular one can rent a second-hand
server that's
already got 4 x 6TB disks (which I mount as RAID, obvs.)
If you’re squeezing 16TB of storage out of 24TB raw, it implies the use of single parity
RAID. I hope you’ve got good backups :)
3:39 p.m.
Hi,
It does. Backups are the customer's responsibility :-}
I've not considered single parity raid to be risky - what do
other people think?
Do you have a hardware controller with non-volatile / battery backed cache
memory?
Traditionally, that's been the only way to avoid the "RAID-5 write hole"
(where you only write a portion of the stripe before a power failure,
creating an undetectable data error).
In other news, I've just bought a pair of 18TB (wow!!! formatted capacity
inexcess of 16TiB!) enterprise Toshiba drives and they boast their own,
integrated, non-volatile cache that seem to also be suitable for use in a
RAID-5 array.
Best wishes,
@ndy
--
andyjpb(a)ashurst.eu.org
http://www.ashurst.eu.org/
0x7EBA75FF
[1] In particular one can rent a second-hand server
that's
already got 4 x 6TB disks (which I mount as RAID, obvs.)
If you’re squeezing 16TB of storage out of 24TB raw, it
implies the use of single parity RAID. I hope you’ve got good
backups :)
5:14 p.m.
On 14/10/2022 16:39, Andy Bennett via BitFolk Users wrote:
Do you have a hardware controller with non-volatile /
battery backed
cache memory?
Nope.
Traditionally, that's been the only way to avoid
the "RAID-5 write hole"
(where you only write a portion of the stripe before a power failure,
creating an undetectable data error).
Ah....
Time for a new server anyway, I think. Had this one getting on for
three years and it wasn't new then.
8:54 p.m.
Hello,
On Fri, Oct 14, 2022 at 04:39:42PM +0100, Andy Bennett via BitFolk Users wrote:
Do you have a hardware controller with non-volatile /
battery backed cache
memory?
Traditionally, that's been the only way to avoid the "RAID-5 write hole"
(where you only write a portion of the stripe before a power failure,
creating an undetectable data error).
Is the RAID-5 write hole not solvable in software? For a long time,
Linux MD parity RAIDs (so -4, -5, -6) have defaulted to using a
write-through cache which is supposed to solve the write hole:
https://www.kernel.org/doc/Documentation/md/raid5-cache.txt
https://lwn.net/Articles/665299/
I thought that non-volatile cache on RAID controllers was so that
you could safely accelerate writes as otherwise you have to disable
the write cache that is in the HDDs themselves as they won't survive
a power loss (and is also how the software in hardware RAID
controllers closes the parity write hole).
If not using traditional RAID (hardware or software), it depends.
ZFS solved the parity write hole by using copy-on-write:
https://pthree.org/2012/12/05/zfs-administration-part-ii-raidz/
notably btrfs has not yet solved it for the parity-based redundancy
levels, and that is one of the reasons why parity-based redundancy
in btrfs is not recommended even as recently as last week:
https://lore.kernel.org/linux-btrfs/86f8b839-da7f-aa19-d824-06926db13675@gm…
But anyway, I don't think people using MD Linux RAID-5 on a modern
kernel with HDD write caches disabled need to worry about about
write hole, unless they specifically went out of their way to
disable the MD cache.
HDD write caches default to "on", by the way…
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
8:22 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
Hello,
On Fri, Oct 14, 2022 at 04:10:58PM +0100, Dom Latter via BitFolk Users wrote:
I've not considered single parity raid to be risky
- what do other people
think?
I wouldn't do single parity with devices as big as 6TB, especially
not at Hetzner who are very likely using consumer HDDs (not ones
recommended to enterprise/NAS use), and often you're not the first
user of them (i.e. they took them out of a server that was rented by
others previously).
The issue is that if you lose one device and then replace it, it
will take a really long time to read 6TB of data off of the other
devices and during that whole time there is no redundancy at all, so
if there are any further unreadable bits on other devices, the data
is lost.
The risk is a little bit lower if you have been regularly scrubbing
your array as that forces it to read everything and find any
unreadable bits. Most Linux distributions using MD RAID do schedule
this for once per month.
It's probably also acceptable if you can restore from backups
without too much grief.
I get why you people would want to do RAID-5 or ZFS raidz1 when
there's only four devices, as RAID-6 or raidz2 is losing half the
capacity…
Sometimes people try to draw conclusions from the Bit Error Rate
/ Unrecoverable Error Rate of the devices, e.g. Toshiba X300 6TB is
1 per 10¹⁴:
https://storage.toshiba.com/docs/support-docs/X300-SalesSheet_English_Web_r…
So they will say once you've read 10¹⁴ bits (12½TB) off of one of those
drives you could expect 1 bit of that to be unreadable. That starts
to look quite worrying when you need to read 6TB off of three of them to
heal the array each time, with normal read load on top.
The thing is, the UER can't mean that, as observed in practice:
https://louwrentius.com/dont-be-afraid-of-raid.html
http://www.raidtips.com/raid5-ure.aspx
HDDs just aren't as bad as that. So I don't know exactly what UER
means or how we're supposed to think of it. Perhaps it is some sort
of minimum reliability guarantee in that if you read 12½ TB off of
one of those drives and 1 bit of it comes out wrong then that's
still functioning as designed, but if 2 bits are wrong then that is
unacceptable and will be replaced under warranty?
Though even that interpretation has difficulties since you can't
read just 1 bit from an HDD, you can only read a sector, which at
these capacities will be 4KiB, and either it will all read correctly
or it won't (and you'll get an error for all of that sector).
Anyway, my main concern wouldn't be the published error rate but
just the fact that at 100MB/sec sustained write, it would take 2½
days to heal the array during which time there is no redundancy. And
100MB/s is really optimistic for a HDD in a system under other load.
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
9:21 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
Hi,
<snip all the very relevant and good points>
Anyway, my main concern wouldn't be the published
error rate but
just the fact that at 100MB/sec sustained write, it would take 2½
days to heal the array during which time there is no redundancy. And
100MB/s is really optimistic for a HDD in a system under other load.
I was pleasantly surprised to learn that the Toshibas I mentioned up thread
were specced at nearly 300MB/s. In my USB 3.0 dock I was getting a
sustained 160 or 170MB/s on to them, sequential write.
So, somehow, these specs are still improving! My previous disks could only
do about 140MB/s. With the new disks I'm clearly limited by my USB3.0 HDD
dock.
OTOH, on an active server, you're not going to get sequential reads and
writes for long before another process comes in and that'll destroy your
performance. You may still not be IOPs limited, but your rebuild won't get
anywhere near the single-threaded read performance.
So yeah, I'd agree that 100MB/s is very optimistic, but I've also been
shocked just how much HDD performance is still improving.
Best wishes,
@ndy
--
andyjpb(a)ashurst.eu.org
http://www.ashurst.eu.org/
0x7EBA75FF
9:26 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
Hi,
So yeah, I'd agree that 100MB/s is very
optimistic, but I've
also been shocked just how much HDD performance is still
improving.
Of course, my actual point was that the time taken to write a full HDD,
end-to-end, assuming you can do it single threaded, seems to have been
stable at around 12 to 24 hours for a few generations of HDD now.
You still "just" have the logistical problem of whether to take your
machine out of service to do the array rebuild.
Best wishes,
@ndy
--
andyjpb(a)ashurst.eu.org
http://www.ashurst.eu.org/
0x7EBA75FF
10:14 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
On Fri, Oct 14, 2022 at 08:22:20PM +0000, Andy Smith via BitFolk Users wrote:
Anyway, my main concern wouldn't be the published
error rate but
just the fact that at 100MB/sec sustained write, it would take 2½
days to heal the array during which time there is no redundancy. And
100MB/s is really optimistic for a HDD in a system under other load.
Sorry I don't know what I did with my calculator there but obviously
100MB/sec is 8640,000MB a day or 8.64TB/day¹, so it's more like 0.7
days.
I still think 100MB/s is really pushing it for a HDD that is doing
anything else at all, though.
Cheers,
Andy
¹ Using all powers-of-ten for unit prefixes here, as storage vendors do.
--
https://bitfolk.com/ -- No-nonsense VPS hosting
10:19 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
On 14/10/2022 21:22, Andy Smith via BitFolk Users wrote:
I wouldn't do single parity with devices as big as
6TB, especially
not at Hetzner who are very likely using consumer HDDs (not ones
recommended to enterprise/NAS use), and often you're not the first
user of them (i.e. they took them out of a server that was rented by
others previously).
Their business model is to auction off previously-used servers,
it being cheaper to flog it off as-is rather than send a techie
into the warehouse to pull disks out. Hence the availability of
machines with "15 x 4.0 TB Enterprise HDD" at ~90 euros per
month.
FWIW - current machine uses "HGST Ultrastar 7K6000" which appear
to be "proper". However it looks like they have been running
for almost six years. So I am advising my customer to move.
The issue is that if you lose one device and then
replace it, it
will take a really long time to read 6TB of data off of the other
devices and during that whole time there is no redundancy at all, so
if there are any further unreadable bits on other devices, the data
is lost.
Well yes, I have sat through a rebuild and it was not pleasant.
The risk is a little bit lower if you have been
regularly scrubbing
your array as that forces it to read everything and find any
unreadable bits. Most Linux distributions using MD RAID do schedule
this for once per month.
Could that be why it was doing an unexpected rebuild the other day?
Nothing in /etc/cron.monthly.
<snip>
So if I convince my customer to move server to something with more
than, say, seven, spinny things what RAID should I be looking at?
Massive thanks all, in particular Andy Bennett, I have read your emails,
Andy Smith: advising me when it's not even your hosting I am talking
about!
thanks
Dom
11:22 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
Hello,
On Fri, Oct 14, 2022 at 11:19:05PM +0100, Dom Latter via BitFolk Users wrote:
On 14/10/2022 21:22, Andy Smith via BitFolk Users
wrote:
On Linux with MD it is usually in /etc/cron.d/mdadm as it is set for
the first Sunday of the month, which is not expressible in normal
cron language (so it fires a script every Sunday to check if it's
the first Sunday).
That's my experience of Debian and Ubuntu anyway. Might be
different on other distributions.
The risk is a little bit lower if you have been
regularly scrubbing
your array as that forces it to read everything and find any
unreadable bits. Most Linux distributions using MD RAID do schedule
this for once per month.
Could that be why it was doing an unexpected rebuild the other day?
Nothing in /etc/cron.monthly. So if I convince my customer to move server to
something with more
than, say, seven, spinny things what RAID should I be looking at?
I think it's quite a hard sell if they're happy with the risk as it
is, but I did not realise there were already so many drives in use.
I didn't see how many devices there are right now, but if we imagine
there are six then at the point where one dies, in terms of
redundancy you effectively have a five drive RAID-0¹. It starts to
make me a bit nervous because of the risk of ANY of those other
devices encountering an error.
At least if you went to RAID-6 you could lose any two at once, which
vastly reduces the risks and if you have lots of drives it doesn't
hugely decrease the capacity or performance.
As for how many drives to have in each RAID-6, opinions are going to
differ. The more you have the higher the performance and capacity. I
might draw the line at about 10 of them. A lot of people with 24
drive bays might go for two 11 drive RAID-6 arrays striped together.
Or 2x10 if two drives were a mirror for the OS to be installed on,
for example.
Here's some thoughts from people asking that question for ZFS
raidz2, which is kind of similar to RAID-6:
https://www.reddit.com/r/zfs/comments/2hvxd4/whats_your_maximum_number_of_d…
Going to ZFS could be an interesting thing to explore because
although ZFS does tend to perform a bit worse than a simple Linux MD
RAID, you do get a lot of nice features included like checksums for
data integrity and self-healing, and compression. And snapshots,
which may or may not be of interest to you.
You can get back some of the performance by having a lot of RAM for
a cache (only used for reads). If you have room for one or more SSDs
then you can put a second layer of cache on that/them (again only
for reads). That doesn't technically need to be redundant since if
it dies ZFS just stops using it, but the resulting decrease in
performance could be problematic, so some people use more than one
SSD for L2ARC. Finally if you did put at least one SSD in then some
or all of it/them could be dedicated to a "ZFS Intent Log" (ZIL,
SLOG, all interchangeable terms). This is for writes only. Again,
loss of the ZIL device means loss of the performance boost² so some
people do mirror them.
A release of ZFS that's coming soon is also meant to have special
data allocation classes, which basically means that you can have a
separate vdev ("array") for file metadata. The point is to put the
small metadata on something fast, and keep the big stuff on the slow
HDDs.
So, lots of interesting ideas there, that were formerly only in the
reach of huge data warehouses with their proprietary solutions for
tiered storage, but clearly a lot to get your head around and
potentially impractical to spec up and test out when you're just
using what can be rented from Hetzner. Simple solutions may be best
at the end of the day.
Cheers,
Andy
¹ Except worse in performance terms because it needs to calculate
parity to read anything, at that point.
² The only time the ZIL is ever read from is if the server crashes
and the filesystem has to replay any writes that went into the ZIL
but didn't yet make it onto the regular storage. So IF your single
ZIL device died AND your server crashed or had power yanked AT THE
SAME TIME, then it would be possible that some data was written to
ZIL and then lost forever.
--
https://bitfolk.com/ -- No-nonsense VPS hosting
11:29 p.m.
New subject: Single parity redundancy and "large" HDDs (Was Re: Re: Sending emails from Bitfolk VPS)
On Fri, Oct 14, 2022 at 11:22:58PM +0000, Andy Smith via BitFolk Users wrote:
A lot of people with 24 drive bays might go for two 11
drive
RAID-6 arrays striped together. Or 2x10 if two drives were a
mirror for the OS to be installed on, for example.
Sigh. 2x12 or 2x11. 3x8 or 3x7 may even make sense there. Point is,
going for a 24-drive RAID-6 is probably not a good idea!
Cheers,
Andy
30 Sep
30 Sep
8:41 p.m.
Not on my Bitfolk VPS, but I recently moved my main server to a
different company. The old one had no IP6 address, the new one has IP4
and IP6.
When I set up the spf for each of the domains hosted there, I put a typo in.
Instead of making the spf say
v=spf1 a mx ip6:nn00:nn98:nn:nn:0:0:0:1/64 -all
I made it
v=spf1 a mx ipv6:nn00:nn98:nn:nn:0:0:0:1/64 -all
They all have dmarc and dkim set up properly, but the effect of that
typo in the spf meant that the ip6 address was invalid, so Google
bounced all mail to gmail addresses, and various other emails just
weren't delivered.
As soon as I put the spf right, all was well.
--
Iain
8:48 p.m.
On Fri, 30 Sep 2022, at 9:13 PM, Ian via BitFolk Users wrote:
.. can you still get emails delivered from Bitfolk
VPSes to the three reliably to inboxes?
Yes, I have all root mail from my VPS delivered to my own email address. I've been
with Fastmail since the beginning of this year but was previously with Microsoft 365
Exchange Online. I had no issues once I had DKIM and SPF configured correctly. I also made
sure that any Bitfolk IP addresses I was using for mail delivery (both IPv4 and IPv6) had
PTR records that matched the hostname - i.e. the IP addresses in the A and AAAA records
resolved back to the hostnames.
Before I had done all these, emails just went straight into the Spam/Junk folder.
Cheers,
Mike
9 p.m.
I’m not sending my mails from bitfolk, but from my own IPs, and for what
I’m aware of, gmail and yahoo are reachable once SPF and DKIM is
correct.
But for MS, it’s impossible. I’m a the point where I stopped trying, I’m
just not reaching them whatever I do. So when I have to write to someone
hosted by them, I ask for a different email address.
--
Alarig
1 Oct
1 Oct
6:14 a.m.
Alarig Le Lay:
But for MS, it’s impossible. I’m a the point where I stopped trying, I’m
just not reaching them whatever I do. So when I have
to write to someone
hosted by them, I ask for a different email address.
The ironic thing is that every non-spambot that gets rejected by me for not
following the RFC rules - eg by not trying again when my end says 'not at
the moment, try again' - runs MS Exchange.
Ian
6:49 a.m.
Alarig Le Lay via BitFolk Users <users(a)mailman.bitfolk.com> writes:
I’m not sending my mails from bitfolk, but from my own
IPs, and for what
I’m aware of, gmail and yahoo are reachable once SPF and DKIM is
correct.
But for MS, it’s impossible. I’m a the point where I stopped trying, I’m
just not reaching them whatever I do. So when I have to write to someone
hosted by them, I ask for a different email address.
I got banned by them once for trying to connect to the bounce address of
a spammer and it took a bunch of waiting and hoops to get unbanned. In
the meantime, I setup my mailserver to deliver MS email through my
work's email server. You could do the same with a commercial
provider. Hosting your own mail server gives you privacy and control,
but going thought an intermediate outgoing server just for MS gives up
only a little and may be worth it.
7:03 a.m.
On Sat 01 Oct 2022 02:49:52 GMT, Ian Kelling wrote:
Alarig Le Lay via BitFolk Users <users(a)mailman.bitfolk.com> writes:
Oh, I’m pretty sure that my work’s email servers don’t have issues with
MS, customers are paying for that. But I don’t want to do any effort for
them, I did my part and everything works elsewhere, so now it’s not my
problem anymore.
I’m not sending my mails from bitfolk, but from
my own IPs, and for what
I’m aware of, gmail and yahoo are reachable once SPF and DKIM is
correct.
But for MS, it’s impossible. I’m a the point where I stopped trying, I’m
just not reaching them whatever I do. So when I have to write to someone
hosted by them, I ask for a different email address.
I got banned by them once for trying to connect to the bounce address of
a spammer and it took a bunch of waiting and hoops to get unbanned. In
the meantime, I setup my mailserver to deliver MS email through my
work's email server. You could do the same with a commercial
provider. Hosting your own mail server gives you privacy and control,
but going thought an intermediate outgoing server just for MS gives up
only a little and may be worth it.
799
days inactive
813
days old
22 comments
9 participants
participants (9)
-
Alarig Le Lay -
Andy Bennett -
Andy Smith -
Anthony Newman -
Dom Latter -
Iain Harrison -
Ian -
Ian Kelling -
Mike Zanker