Category computers

Anyone want to see a bug in Mail.app v2.1 (752/752.2)? Here: mail_imap_single_subfolder_bug.mp4

In short, I have two folders under INBOX (in this case .Logs and .Other) which aren’t defined as special folders (.Sent, .Drafts, .Junk, and .Trash are being used as special folders in Mail.app). If I delete one of them, in this case .Other, suddenly .Logs is no longer visible.

In short, if there is one folder besides the ‘special’ ones off of INBOX, it won’t be shown. And yes, I’ve confirmed that the folder is subscribed. Pointing SquirrelMail or Thunderbird at the account shows all folders as expected.

This is with Courier-IMAP 4.1.1 running on FreeBSD 6.1.

And yes, this has been reported to Apple as Problem ID 4764544.

Can any of you tell me why the following matches no files (and thusly won’t back things up?

tar -czP --include '*.sh' -f /home/telematic/backup/`date +"%Y-%m-%d_scripts.tar.gz"` /usr/local/sbin

This is with BSD tar, and there are files matching *.sh under there. In fact:

root@rowla:~telematic/backup# find /usr/local/sbin -name "*.sh"
/usr/local/sbin/fastcgi/fastcgi-katester.sh
/usr/local/sbin/fastcgi/fastcgi-c0nsumer.sh
/usr/local/sbin/fastcgi/fastcgi-rekram.sh
/usr/local/sbin/fastcgi/fastcgi-sithspawn.sh
/usr/local/sbin/fastcgi/fastcgi-nuxxfriends.sh
/usr/local/sbin/fastcgi/fastcgi-tnorman.sh
/usr/local/sbin/fastcgi/fastcgi-secure.sh
/usr/local/sbin/automysqlbackup.sh
/usr/local/sbin/webalizer_daily.sh
/usr/local/sbin/dailybackup.sh
/usr/local/sbin/lighttpd-hash.sh
root@rowla:~telematic/backup#

So, uhm… What am I doing wrong?

DreamHost Cancelation
(Click for full res image…)

In case any of you are interested, the process for canceling an account with DreamHost is completely automated. In fact, it’s easier than signing up for the account in the first place. When canceling the site simply calculates whether or not it has been more than 97 days since signing up for the account and if not, automatically supplies a refund.

Maybe they get quite a few of these or something.

Anyway, I’m (obviously) going to wait a few days to be sure things go through as expected with the credit card company, but this is definitely a bright spot in this whole ordeal with DreamHost.

Oh, and yes, all sites are now on rowla.nuxx.net at the wonderful host which is Waveform Technology Inc, LLC. As long as I can provide a stable, unhacked box for another more than two years I’ll be content.

So it seems that my choice of webserver for rowla.nuxx.net, lighttpd, isn’t working out so well. It seems that the Nokia 6600 (and I’d imagine others), as part of implementing the Nokia Image Upload Server API v1.1 sends an Expect: 100-continue, but lighttpd always returns 417 Expectation Failed. As can be seen here, the ‘100 – Continue’ part of HTTP/1.1 simply hasn’t been implemented.

A bug report (#377 – Incorrect handling of the 100 (Continue) Status) has been submitted, but it looks like there hasn’t been any work done on it in quite a while.

I presume that Nokia uses that because then the client will only send another chunk of image after the server has received the last, helping ensure data doesn’t get lost somewhere in the high latency / flaky world of GPRS. Or maybe because they anticipate the Image Upload feature of the phones to be used on potentially overloaded ISP / telco servers and they don’t want the phones sending data faster than the server can handle it.

As is stated in RFC2068, 10.1.1 – 100 Continue:

The client may continue with its request. This interim response is used to inform the client that the initial part of the request has been received and has not yet been rejected by the server. The client SHOULD continue by sending the remainder of the request or, if the request has already been completed, ignore this response. The server MUST send a final response after the request has been completed.

I don’t think I can use some sort of proxy for it either, because it seems that Expect headers need to be forwarded by proxies. (RFC2616, 14.20)

So, back to Apache, or I can try to find another way around this… At least I know what’s causing it. Until then, no from-phone updating of my moblog can occur.

rowla.nuxx.net lives!

PING 204.11.33.41 (204.11.33.41): 56 data bytes
64 bytes from 204.11.33.41: icmp_seq=0 ttl=42 time=44.605 ms
64 bytes from 204.11.33.41: icmp_seq=1 ttl=42 time=34.532 ms
64 bytes from 204.11.33.41: icmp_seq=2 ttl=42 time=31.647 ms
64 bytes from 204.11.33.41: icmp_seq=3 ttl=42 time=37.469 ms
--- 204.11.33.41 ping statistics ---
4 packets transmitted, 4 packets received, 0.0% packet loss
round-trip min/avg/max/std-dev = 31.647/37.063/44.605/4.818 ms

I don’t completely understand how / why yet, nor how I’ll manage it, but I’ve got Postfix authenticating against mysql and pulling the virtual user / domain from there. I think that’s enough work for tonight.

Tomorrow the server goes into colocation.

On a suggestion I took a look at using lighttpd instead of Apache. Well, from what I can see for now it looks like an excellent substitute, especially for my needs. As long as I can get its mod_rewrite working as I need, it’ll be exactly what I need.

(Crap, just thought of one thing… Gallery doesn’t know how to deal with it, I don’t think… Must investigate it more… May need Apache. hrm.)

Anyway, here’s another wonderful error I received from DreamHost. Enjoy:

attempt to access field 'dbname' in object of class Ndn::Dreamhost::QueryTime at /usr/local/lib/site_perl/Ndn/Common/Object.pm line 949
Ndn::Common::Object::AUTOLOAD('Ndn::Dreamhost::QueryTime=HASH(0xc0ecc04)', 'undef') called at /usr/local/ndn/web/dhwebpanel/index.cgi line 117
Apache::ROOTpanel_2edreamhost_2ecom::index_2ecgi::handler('Apache=SCALAR(0xba6f09c)') called at /usr/lib/perl/5.8/Apache/Registry.pm line 149
eval {...} called at /usr/lib/perl/5.8/Apache/Registry.pm line 149
Apache::Registry::handler('Apache=SCALAR(0xba6f09c)') called at /dev/null line 0
eval {...} called at /dev/null line 0

So my new server is having a couple of problems. For one, when I tried to turn it on last night, it took three tries to power up. Now today it’s rebooting whenever there is a lot of disk load. See?

/usr/ports/comms/yawmppp/
/usr/ports/comms/zmtx-zmrx/
/usr/ports/comms/zssh/
/usr/ports/converters/Makefile
rm: /usr/ports/converters/Makefile: Bad file descriptor
Read from remote host 192.168.0.2: Connection timed out
Connection to 192.168.0.2 closed.
-bash-3.1$

(That was while running an update to the ports collection.)

I’m starting to think that my server wasn’t really compromised, and it just happened to be having power problems at the time it got hacked, and the massive amount of disk activity during the mass deletion / replacement caused the server to reboot.

So, I’ve ordered a FSP550-80GLC (PDF drawing) to hopefully resolve this problem. I likely won’t get the server into the colocation facility until next week, but I’ll see what I can do to get it limping along for the time being so that the OS can at least be installed and configured.

jawbreaker:~> time ls
bin gallery2_data logs Maildir mysql_backup www
0.000u 0.000s 0:02.84 0.0% 0+0k 0+0io 158pf+0w
jawbreaker:~>

Anyone want another reason why I’m leaving DreamHost?

That’s just absurd.

I can’t wait to get home and (hopefully) start getting the new server together.