Re: Debian Co-location in USA
On Tue, Jul 15, 2003 at 01:27:04PM +1000, Jeremy Lunn wrote: > One of my clients is looking into hosting a server in the USA, since it > costs much less than doing it here in Australia. For Debian, you have a few options. I hear a lot of good things about Dream Host ( http://www.dreamhost.com ) and they're Debian-based. You can also simply convert a Red Hat installation into a Debian installation. It's a little nervous-making, but doable! With most hosting providers, you can choose the partitioning scheme, so just have a small partition prepared where you can untar an installed system to, chroot to that and run lilo, then make it restart and repurpose the other volumes as you please. I have done this on a server where custom partitioning wasn't an option as well. I simply used the swap partition. I converted the swap to an ext2 filesystem, dropped in a tiny Debian install, restarted with that as the new root and ran fsck from the small install to lay things out the way I wanted on the rest of the drive. When I was done, that tiny partition just became my swap again. :-) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Exim4 and Reverse Lookup Delay
Hello everyone, i have a problem with my exim4-4.20. The installed configuration is exim4-4.20 and bind9 on woody. The machine is hosted by a big german hosting company. Now i want to send mails via this exim4 smtp. But anytime i connect exim4 needs at least 90s to accept mails. I guess that´s the time needed to make a reverse lookup. Is there any possiblity to have a wrong bind configuration which leads to this problem ? (If i try the same connect from my university account it needs only 1 second ) Thanks in advance, Rüdiger Pryss -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
BIND 8 or 9 version ?
hi2all please - tell me something about preferred vesion of BIND: 8 or 9 ? Is 9 version more better than 8 ? Should I to upgrade from my 8.3 up to 9 ? Debian 3.0_r1 thx2all -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
IBM Ultrium Model 3580
I have a Plasmon LTO 10.1 SCSI Tape Library, running with a IBM Ultrium Model 3580 tape drive, which I have the honor of configuring under Linux. While it may be more appropriate of me to post this question to a SCSI forum, I wanted to gain the insight of those here who may have worked with this device in the past. Writing using tar seems unreliable. The first several writes/reads worked fine, but after awhile I get this error when writing to the beginning of tape: Jul 16 13:55:10 ma-bell kernel: ASC=4b ASCQ= 0 Jul 16 13:55:10 ma-bell kernel: Raw sense data:0xf0 0x00 0x0b 0x00 0x00 0x00 0x1e 0x1c 0x00 0x00 0x00 0x00 0x4b 0x00 0x00 0x00 0x10 0x51 0x00 0x00 0x00 0x01 0x20 0x20 0x20 0x20 0x20 0x20 0x20 0x00 0x00 0x00 0x00 0x00 0x00 0x00 Jul 16 13:56:32 ma-bell kernel: st0: Error with sense data: Current st09:00: sns = 70 5 When running 'mt-st reten' I get this error: Jul 16 13:56:32 ma-bell kernel: ASC=24 ASCQ= 0 Jul 16 13:56:32 ma-bell kernel: Raw sense data:0x70 0x00 0x05 0x00 0x00 0x00 0x00 0x1c 0x00 0x00 0x00 0x00 0x24 0x00 0x00 0xc8 0x00 0x01 0x00 0x00 0x00 0x01 0x20 0x20 0x20 0x20 0x20 0x20 0x20 0x00 0x00 0x00 0x00 0x00 0x00 0x00 The software page of the plasmon only mentions vendors such as Arkeia and Computer Associates for supported software. I am not so interested in getting the Robot working as much as just having the ability to write to tape. I greatly appreciate any advise. Could anyone also recommend a good SCSI-related mail list or newsgroup in which this question would be better suited? thanks, hank -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: woody + proftpd + ldap = segfault
sorry - I should have sent this to the list instead of just you - did
this help?
Just out of interest and because this helped me once - try apt-getting
nscd and start it and see if it works then.
Secondly, do you know your LDAP is working? Is it using nsswitch and
PAM?
Michael Moritz
Rémi Letot wrote:
Eduard Ballester <[EMAIL PROTECTED]> writes:
hi
swoog (192.168.50.11[192.168.50.11]) - ProFTPD terminating (signal 11)
Does it ring a bell ?
There is no sufficient information, perhaps you must up the debug
level (try -d 9). If you can't see the error, try to run it without
mod_ldap configuration (using your system users).
upping the debug level doesn't provide anything more (actually it
seems 5 is the maximum debug level)
working without ldap works fine
working with ldap+tls doesn't work because the ldap server can't speak
tls, but in that case proftpd doesn't die. It simply answers 'no such
user'.
If nothing works you can use a strace command (system call tracer) to
obtain more data. The most simple use is strace -p
As proftpd spawns a new daemon for each incomming connection, I did an
strace -fn $proftpdpid, and here is what comes out when I try to
connect : (snipped the former lines as it's quite long)
-<
[pid 23172] read(5, "127.0.0.1\tswoog\tlocalhost\n192.16"..., 4096) =
312
[pid 23172] read(5, "", 4096) = 0
[pid 23172] close(5)= 0
[pid 23172] munmap(0x40015000, 4096)= 0
[pid 23172] socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 5
[pid 23172] connect(5, {sin_family=AF_INET, sin_port=htons(53),
sin_addr=inet_addr("192.168.60.5")}}, 28) = 0
[pid 23172] send(5,
"\226\271\1\0\0\1\0\0\0\0\0\0\tlocalhost\0\0\34\0\1", 27, 0) = 27
[pid 23172] gettimeofday({1057942581, 978654}, NULL) = 0
[pid 23172] poll([{fd=5, events=POLLIN, revents=POLLIN}], 1, 5000) = 1
[pid 23172] recvfrom(5,
"\226\271\205\200\0\1\0\0\0\1\0\0\tlocalhost\0\0\34\0\1"..., 1024, 0,
{sin_family=AF_INET, sin_port=htons(53),
sin_addr=inet_addr("192.168.60.5")}}, [16]) = 68
[pid 23172] close(5)= 0
[pid 23172] socket(PF_UNIX, SOCK_STREAM, 0) = 5
[pid 23172] connect(5, {sin_family=AF_UNIX,
path="/var/run/.nscd_socket"}, 110) = -1 ENOENT (No such file or
directory)
[pid 23172] close(5)= 0
[pid 23172] open("/etc/hosts", O_RDONLY) = 5
[pid 23172] fcntl64(5, F_GETFD) = 0
[pid 23172] fcntl64(5, F_SETFD, FD_CLOEXEC) = 0
[pid 23172] fstat64(5, {st_mode=S_IFREG|0644, st_size=312, ...}) = 0
[pid 23172] old_mmap(NULL, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40015000
[pid 23172] read(5, "127.0.0.1\tswoog\tlocalhost\n192.16"..., 4096) =
312
[pid 23172] read(5, "", 4096) = 0
[pid 23172] close(5)= 0
[pid 23172] munmap(0x40015000, 4096)= 0
[pid 23172] socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 5
[pid 23172] fcntl64(5, F_GETFL) = 0x2 (flags O_RDWR)
[pid 23172] fcntl64(5, F_SETFL, O_RDWR|O_NONBLOCK) = 0
[pid 23172] connect(5, {sin_family=AF_INET, sin_port=htons(389),
sin_addr=inet_addr("127.0.0.1")}}, 16) = -1 EINPROGRESS (Operation now
in progress)
[pid 23172] select(1024, NULL, [5], NULL, NULL) = 1 (out [5])
[pid 23172] getpeername(5, {sin_family=AF_INET, sin_port=htons(389),
sin_addr=inet_addr("127.0.0.1")}}, [16]) = 0
[pid 23172] fcntl64(5, F_GETFL) = 0x802 (flags
O_RDWR|O_NONBLOCK)
[pid 23172] fcntl64(5, F_SETFL, O_RDWR) = 0
[pid 23172] getpeername(5, {sin_family=AF_INET, sin_port=htons(389),
sin_addr=inet_addr("127.0.0.1")}}, [16]) = 0
[pid 23172] brk(0x80af000) = 0x80af000
[pid 23172] uname({sys="Linux", node="swoog", ...}) = 0
[pid 23172] time(NULL) = 1057942581
[pid 23172] write(5, "0\24\2\1\1`\17\2\1\2\4\10o=ispman\200\0", 22) =
22
[pid 23172] select(1024, [5], [], NULL, NULL) = 1 (in [5])
[pid 23172] read(5, "0\f\2\1\1a\7\n\1\0\4\0\4\0", 16384) = 14
[pid 23172] time(NULL) = 1057942581
[pid 23172] --- SIGSEGV (Segmentation fault) ---
[pid 23172] rt_sigaction(SIGSEGV, {SIG_DFL}, {0x804f4ac, [SEGV],
SA_RESTART|0x400}, 8) = 0
[pid 23172] write(2, "swoog (192.168.50.11[192.168.50."..., 71) = 71
[pid 23172] _exit(1)= ?
-<
And then the parrent process spawns some more lines before it stays
still waiting for a new connection.
The sig11 is clearly there, but I'm not a developper, so can someone
interpret this? Or is it still too scarse in informations? If more
info is needed, how can I provide it ?
I have tried to recompile debian's package with the latest release of
mod_ldap, but it displays the same behaviour.
Thanks for your help,
RE: Exim Reverse Lookup
Rudiger, Mein Deutsch ist sehr slecht, aber: Die Debian mailinglist ist ein Engelse sprache mailinglist. In English: the debian mailinglists are all in English, unless stated otherwise. See www.debian.org for details. Maybe there are also German language ones, check this out on www.debian.org. For now: please state your question in English so we can understand your problem :) Pim - PingWings - - - - - - - - - - - - - - - - - - - - - M: 06-28474259 E: [EMAIL PROTECTED] I: www.pingwings.nl - -Original Message- From: Ruediger Pryss - Elitec.Info [mailto:[EMAIL PROTECTED] Sent: donderdag 17 juli 2003 2:08 To: [EMAIL PROTECTED] Subject: Exim Reverse Lookup Hallo zusammen, ich habe folgende Frage. Ich habe einen root-Server von 1und1. Auf dieser Kiste läuft debian woody mit einem exim 4.20. Nun zum Problem. Ein Connect zum exim dauert ca. 90 Sekunden, ich vermute einfach mal, dass dies die Zeit ist, die für den reverse lookup gebraucht wird. Diese Verzögerung tritt allerdings nur auf, wenn ich mich von meinem Heimarbeitsplatz über T-Online einwähle. Telnete ich exim z.b. über einen Uni- zugang an, dauert das connecten gerade mal 1s. Den Mail- server von 1und1 erreiche ich ja auch mit meinem T-Online Zugang nach dieser knappen Sekunde. Was könnte ich nun eventuell anders einstellen, so dass dieser lookup konstant 1 Sekunde beträgt, oder anders gefragt, was habe ich eventuell auf dem root-Server falsch eingestellt. Es läuft ein bind9 und 1und1 stellt den Secondary DNS. Gruß, Rüdiger -- Elitec Computersysteme Fon 07322 / 9333444 Braunenbergweg 8Fax 07322 / 958794 89537 Giengen mobil 0173 / 8734229 -- Internet www.elitec.infomail [EMAIL PROTECTED] -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
