Re: [gnu.org #1441879] [fencepost] /com/archive/planet is not usable

[Sylvain Beucler]

Hi Ian,

Thanks for the mailing list - testing it with this very mail.

The old archive is not readable anymore though:
$ ls -lh /com/archive/planet
-rw------- 1 Debian-exim Debian-exim 521K Nov 13 12:18 /com/archive/planet

Cheers!
Sylvain

On 13/11/2019 22:30, Ian Kelling via RT wrote:
> I forgot, out list creation script automatically emails
> the password to a single list owner. I did it, you can share
> it with the other list admin. The list should be functional within about
> 10 minutes. I'm replacing /com/archive/planet as you suggested.
>
> On Wed Nov 13 15:54:05 2019, beuc wrote:
>> Hi Ian,
>>
>> - Let's move address@hidden to a mailman list.
>> - Jose and I as list admins is OK.
>> - My GPG key is correct; in doubt you can store the password in a file
>> in fencepost homedir.
>> - The new mailing list is public.
>> - The past archive cannot be imported: it is semi-private, as it was
>> readable only by people with a fencepost account, so it's a privacy
>> setup we can't replicate with mailman.
>> - The filtered archive at /home/b/beuc/planet/archive.mbox can be used
>> to replace the 3GB /com/archive/planet if you wish so.
>>
>> Cheers!
>> Sylvain
>>
>> On 13/11/2019 18:18, Ian Kelling via RT wrote:
>>> I think moving address@hidden to become a mailman list would be the
>>> best
>>> option. We have a good spam filtering system there for new messages,
>>> the
>>> old ones, hopefully your manual cleaning is good enough. Do you want
>>> it
>>> as a private list or a public one? I can do the creation and
>>> migration. Let me know what you think.
>>>
>>> After I do it, I will send you the list admin password. Should I also
>>> send it to jemarch? Anyone else?
>>>
>>> I have this as your most recent gpg key is that right?
>>> 4227 3C1A E37F C434 7CF0  7912 8FF1 CB6E 8D89 059F
>>>
>>> For jemarch, i have
>>>
>>> BDFA 5717 FC1D D35C 2C38  32A2 3EF9 0523 B304 AF08
>>>
>>>
>>> On Sat Nov 02 13:54:35 2019, beuc wrote:
>>>> Hello sysadmins,
>>>>
>>>> When can we initiate conversation about this?
>>>>
>>>> I also note that large messages (or possibly messages with
>>>> attachments)
>>>> do not get seem to be processed by the mail filters, which explains
>>>> why
>>>> we currently have such a large archive with lots of large
>>>> unprocessed spam.
>>>>
>>>> (btw I did some spam/size/manual filtering: there's 0.5MB/3000MB and
>>>> 150/40000 messages of legitimate content in the archive!)
>>>>
>>>> Another option would be to move address@hidden to a mailman archive,
>>>> thoughts?
>>>>
>>>> Cheers!
>>>> Sylvain
>>>>
>>>> On 22/10/2019 10:52, Sylvain Beucler wrote:
>>>>> Hello,
>>>>>
>>>>> [resending as I didn't get a ticket, I suppose discussing /mail
>>>>> filtering/ triggered some /filtering/ measures.]
>>>>>
>>>>> /com/archive/planet currently weighs 3GB, with ~40000 entries:
>>>>> $ ls -lh /com/archive/planet
>>>>> -rw-r--r-- 1 Debian-exim Debian-exim 3.0G Oct 20 04:12
>>>>> /com/archive/planet
>>>>>
>>>>> [snip!]
>>>>>
>>>>> In the current state, the archive is near-unusable and sheds
>>>>> opacity on
>>>>> GNU Planet's maintenance.
>>>>>
>>>>>
>>>>> Is there a way to filter out, for past and future messages, e-mails
>>>>> that
>>>>> trigger a high level of UBM?
>>>>> Maybe a cut-off at 5 would be reasonable, what do you think?
>>>>>
>>>>> [snip!]