|
Grex > Oldcoop > #294: Why Grex lost its mail partition | |
|
| Author |
Message |
| 25 new of 176 responses total. |
ric
|
|
response 92 of 176:
| 
Dec 1 20:29 UTC 2005 |
I care about Grex and M-Net (for different reasons).
And I still think that anyone who uses either system with the expectation that
their files are safe OR secure is a fool, and I don't have any sympathy for
people who lost important email they had stored on grex.
|
mcnally
|
|
response 93 of 176:
|
Dec 1 20:35 UTC 2005 |
Won't ric be surprised when he finds out I used my staff access to
delete his home directory, conference participation files, and uid!
Just kidding, of course, but if he thinks users shouldn't expect their
e-mail to be safe from sudden disappearance I'm not sure what else on
the system ought to be sacrosanct..
|
nharmon
|
|
response 94 of 176:
|
Dec 1 20:38 UTC 2005 |
If most of the users agree with you mcnally, then that should be one of
Grex's goals.
|
krj
|
|
response 95 of 176:
|
Dec 1 20:57 UTC 2005 |
Mike in resp:91 :: before Grex left the Pumpkin, there were numerous
times when I dropped Steve off there after we got back from work,
and he worked on Grex for some hours in the very late evening or
early morning.
|
glenda
|
|
response 96 of 176:
|
Dec 2 02:14 UTC 2005 |
I seem to remember a few times that you dropped him off at the Pumpkin when
you got back into town, and picked him up there in the morning to go back to
work.
For those advocating having an equivalent system for doing upgrades and
recoveries: where do we store it? The colo charges for space. If a staffer
stores it we still have problems with access unless that staffer is the ONE
doing the upgrade/recovery.
|
ric
|
|
response 97 of 176:
|
Dec 2 15:22 UTC 2005 |
re 93 - I would be surprised, and I'd probably ask for your removal if you
did that on purpose without good reason. But it wouldn't really bother me
much. I'd just create a new account. I participate in two conferences - coop
and agora. And I have used the forget statement on all but one item in agora.
I don't have any files in my home directory that are important to me.
The only thing that might upset me is if I was unable to get my username "ric"
back, since I'm pretty much been known as "ric" in the mnet/grex world since
1986.
(Though I think there was a period of time in the mid 90s where someone else
had that ID on Grex cuz I got reaped)
|
slynne
|
|
response 98 of 176:
|
Dec 3 00:10 UTC 2005 |
Even though we are just a small organization, there is nothing wrong
with us doing the best we can in all situations. I also think that
criticism is ok although I sometimes think that some people around here
have trouble presenting their criticism in the best possible way. It is
pretty easy to start feeling defensive about things.
As for the email loss. It was a mistake. It cant be undone and that is
that. No one did it out of malice. And even the most competant
technical people make mistakes sometimes and email sometimes gets lost
even at for-profit firms.
As for what we can do to prevent such a loss in the future...Well,
there are a lot of good ideas being presented here. I dont know what
the answer is. Our finances arent great and I know that there is a
reluctance to spend a lot of money. However, exploring backup options
is really something we should do.
|
keesan
|
|
response 99 of 176:
|
Dec 3 05:06 UTC 2005 |
I just found the info that I had saved in a recent email and it is actually
nice not to have to go through all 200 or so old mails deciding if there was
anything important in them, so I am actually grateful now, and pine starts
up so much faster with an empty inbox. I wish spamassassin would work again.
|
bhoward
|
|
response 100 of 176:
|
Dec 3 07:56 UTC 2005 |
Have you tried in the last few days. I reinstalled spamassassin and spamd
a day or so ago.
|
keesan
|
|
response 101 of 176:
|
Dec 3 14:41 UTC 2005 |
I had been using a copy in someone else's account, because he said he updated
it more often. I will switch to the grex version, thanks. I had gone back
to my old filter, which is about 10 pages long and lets some things through.
|
tsty
|
|
response 102 of 176:
|
Dec 4 07:35 UTC 2005 |
re #79 ... excuse me! it wouild seem, that *i* undrsatnd "the nature of the
upgrade," one hulluva lot better than either you or other staff or other borg!
shit!
"there was no path back" --- that is *precisely* the sysadmin situation
for which i have been * t r a i n e d * !!
whtether it is an air defense missle ssytem or a fscking os upgrade - the
cover-yur-ass attributes are identical.
somewhare along the line i copied this:
Worse, a great deal of the delay was because we as staff really failed to work
together effectively. We ran into deep differences in basic philosophy about
how grex should be run that cost us extra days. Because we didn't all agree
on what we were going to be doing before we started, our preparation for the
rebuild was not complete. We ended up redoing significant portions of the
job more than once.
i don t know, at this moment, where it came from, but i did *not*write it.
some rooty-tooty (not sTeve) did .. and borg & staff are imtimately
responsible for the fsxk-up.
mostyl borg!
'in-place' .... WTF!
|
cross
|
|
response 103 of 176:
|
Dec 4 15:29 UTC 2005 |
I believe that Jan said that.
|
naftee
|
|
response 104 of 176:
|
Dec 4 16:33 UTC 2005 |
StEve
steVE
sTeve
STeVE
|
cross
|
|
response 105 of 176:
|
Dec 4 17:35 UTC 2005 |
Quick! What's 5 choose 2? Answer: (5!)/((5-2)!(2)!) = (5!)/((3!)(2!)) =
(5*4)/2 = 5*2 = 10. Think of the permutations of the capitalization of
letters in Steve's name this way: Given a string of 5 characters, taken
from the Alphabet {0, 1}, how many ways may I write such a string with
exactly two 1's? Clearly there are 5 choose 2 such ways, and as we have
seen, that means 10 possibilities. Now, I take a 1 to mean a capital
letter and a 0 to be a lowercase letter and enumerate:
STeve = 11000
StEve = 10100
SteVe = 10010
StevE = 10001
sTEve = 01100
sTeVe = 01010
sTevE = 01001
stEVe = 00110
stEvE = 00101
steVE = 00011
These make up the set of permutations of Steve's name with his preferred
number of capitals (though his preferred choice is one specific element).
|
glenda
|
|
response 106 of 176:
|
Dec 4 21:22 UTC 2005 |
His preferred choice came about by accident. When he first started using
conferencing systems, he didn't release the shift key fast enough. He went
to the National Computer Conference and while in a conversation someone asked
him if he was the S T eve. He laughed and replied that he was and they had
a great time talking. He decided that it was a good thing to keep and has
used it, purposely, even since.
|
scholar
|
|
response 107 of 176:
|
Dec 4 21:57 UTC 2005 |
I purposely pervert his choice by being a wiseguy and taking the other oddity
of his name ('), which is applied to his last name, and applying it to his
first name.
I didn't do this because I had a great time talking.
|
naftee
|
|
response 108 of 176:
|
Dec 4 22:49 UTC 2005 |
re 105 I prefer using my calculator to solve those types of problems, but
really; i was just goofing around !@
|
sholmes
|
|
response 109 of 176:
|
Dec 5 08:19 UTC 2005 |
that's also the number of handshakes in a party with 5 ppl, if everyone shakes
handswith everyone else.
|
cross
|
|
response 110 of 176:
|
Dec 5 17:57 UTC 2005 |
That's true. Think of each bit as being two people shaking hands.
|
janc
|
|
response 111 of 176:
|
Dec 5 21:50 UTC 2005 |
I agree 100% that we shouldn't have rebuilt the system by overwriting
the old disk partitions. One of the recommendations I made in my
post-mortem item immediately after the new system came up was to never
do that again. Alas, I did not make that recommendation before the
rebuild - though that was certainly part of the upgrade method defined
in Grexdoc - that's why the ALT partitions exist. But I'm really not an
experienced system adminstrator anyway. I'm not sure that the need to
avoid a destructive rebuilt was as clear in my head before this fiasco
as it was afterwards. Live and learn. In any case, I wasn't around to
give any recommendations.
Before the upgrade, John was really the only active staff member. He
was doing the reboots. He was debugging grexdoc on another machine. He
was reluctant to undertake the rebuild by himself though. My impression
was that there was something of a panic at the board meeting. Grex was
crashing regularly, and their wasn't much of staff plan to do anything
about it. STeve, a board member and a staff member, responded to the
emergency by committing his next weekend to a Grex upgrade.
I had been neglecting Grex so completely that I didn't even know about
it until I talked to John and Mary on the Grex walk the morning before
the upgrade. There was never really any staff meeting to discuss the
upgrade. If there had been, we might have given it enough thought to
realize that there were alternatives to doing a destructive rebuild. In
fact, I think we have a spare (rebuilt) 18G drive laying around. I
think with that we could have managed the rebuild without buying a new
disk. But buying a disk would have made sense too. We rushed into the
upgrade. It felt like Grex was in crisis. If we had held a staff
meeting first, I'm not sure anyone except John would have shown up.
|
tod
|
|
response 112 of 176:
|
Dec 5 22:45 UTC 2005 |
Tod said the board should be trying harder to get more staff. Well, I'm not
on the board right now, but I think I speak for them when I say, they're
open to suggestions.
Could have fooled me. I see nothing but excuses being made and "we do/did
ENOUGH already"
Excuse me for asking for something more than an MOTD, decent backup, and
effort to find staff with more availability. How dare me for making
suggestions. Shame shame.
|
naftee
|
|
response 113 of 176:
|
Dec 5 23:07 UTC 2005 |
shame on you.
|
tod
|
|
response 114 of 176:
|
Dec 5 23:50 UTC 2005 |
THanks Michael Moore!
|
naftee
|
|
response 115 of 176:
|
Dec 6 00:23 UTC 2005 |
thanks tod :L)(
|
ric
|
|
response 116 of 176:
|
Dec 6 15:03 UTC 2005 |
re 100 - I'm not suggesting that staff doesn't do the best they possibly can
to avoid email loss and other such things. I'm suggesting that we as users
should not expect or demand anything more. The fact is, if this were a
commercial organization, there would be daily tape backups, stored off site,
our hardware would probably be more "enterprise" level and all sorts of such
things - policies in place to prevent such occurences, and paid employees
whose PRIMARY responsibility is maintenance of the server(s).
Grex is nobody's primary responsibility. I'm pretty sure it's nobody's
secondary responsibility - at the very best, I would expect Grex to come
somewhere after job and family.
|