2

Note: I've replaced IP addresses, database names, and server users with examples. That shouldn't affect anything.


The Setup

I've installed unixODBC (yum install unixODBC) and IBM's official iSeries ODBC driver (yum install ibm-iaccess-1.1.0.5-1.0.x86_64.rpm, RPM downloaded from IBM's logged-in areas). The installation successfully added the drivers to /etc/odbcinst.ini:

[IBM i Access ODBC Driver]
Description             = IBM i Access for Linux ODBC Driver
Driver          = /opt/ibm/iaccess/lib/libcwbodbc.so
Setup           = /opt/ibm/iaccess/lib/libcwbodbcs.so
Driver64                = /opt/ibm/iaccess/lib64/libcwbodbc.so
Setup64         = /opt/ibm/iaccess/lib64/libcwbodbcs.so
Threading               = 0
DontDLClose             = 1
UsageCount              = 1

[IBM i Access ODBC Driver 64-bit]
Description             = IBM i Access for Linux 64-bit ODBC Driver
Driver          = /opt/ibm/iaccess/lib64/libcwbodbc.so
Setup           = /opt/ibm/iaccess/lib64/libcwbodbcs.so
Threading               = 0
DontDLClose             = 1
UsageCount              = 1

The referenced library files do exist, and they are correctly linked (checked via ldd, no missing links).

My ~/.odbc.ini file looks like this:

[Foo]
Driver          = IBM i Access ODBC Driver
DATABASE        = FooDB
SYSTEM          = 123.45.67.8
HOSTNAME        = 123.45.67.8
PORT            = 446
PROTOCOL        = TCPIP

The Problem

When I run isql Foo USER PASSWORD -v, I get this output after roughly a minute or so:

user@example.com [~]# isql FooDB USER PASSWORD -v
[08S01][unixODBC]
[ISQL]ERROR: Could not SQLConnect

Troubleshooting

Sounds like it's timing out, right?

ping 123.45.67.8 returns:

user@example.com [~]# ping 123.45.67.8
PING 123.45.67.8 (123.45.67.8) 56(84) bytes of data.
64 bytes from 123.45.67.8: icmp_seq=1 ttl=63 time=29.8 ms
64 bytes from 123.45.67.8: icmp_seq=2 ttl=63 time=29.8 ms
64 bytes from 123.45.67.8: icmp_seq=3 ttl=63 time=29.8 ms
64 bytes from 123.45.67.8: icmp_seq=4 ttl=63 time=31.0 ms
64 bytes from 123.45.67.8: icmp_seq=5 ttl=63 time=29.9 ms

telnet 123.45.67.8 446 returns:

user@example.com [~]# telnet 123.45.67.8 446
Trying 123.45.67.8...
Connected to 123.45.67.8.
Escape character is '^]'.
foobar
^]
telnet> quit
Connection closed.

Enabling the ODBC log with Trace and TraceFile in /etc/odbcinst.ini produces output like this:

[ODBC][22093][1454628360.104274][__handles.c][450]
                Exit:[SQL_SUCCESS]
                        Environment = 0x13a4750
[ODBC][22093][1454628360.104316][SQLAllocHandle.c][364]
                Entry:
                        Handle Type = 2
                        Input Handle = 0x13a4750
[ODBC][22093][1454628360.104339][SQLAllocHandle.c][482]
                Exit:[SQL_SUCCESS]
                        Output Handle = 0x13a5080
[ODBC][22093][1454628360.104363][SQLConnect.c][3614]
                Entry:
                        Connection = 0x13a5080
                        Server Name = [FooDB][length = 4 (SQL_NTS)]
                        User Name = [USER][length = 7 (SQL_NTS)]
                        Authentication = [*******][length = 7 (SQL_NTS)]
                UNICODE Using encoding ASCII 'ISO8859-1' and UNICODE 'UCS-2LE'

                DIAG [08S01]

[ODBC][22093][1454628423.118602][SQLConnect.c][3982]
                Exit:[SQL_ERROR]
[ODBC][22093][1454628423.118628][SQLError.c][430]
                Entry:
                        Connection = 0x13a5080
                        SQLState = 0x7fff9a5bd7b0
                        Native = 0x7fff9a5bd5a8
                        Message Text = 0x7fff9a5bd5b0
                        Buffer Length = 500
                        Text Len Ptr = 0x7fff9a5bd5ae
[ODBC][22093][1454628423.118656][SQLError.c][467]
                Exit:[SQL_SUCCESS]
                        SQLState = 08S01
                        Native = 0x7fff9a5bd5a8 -> 10060
                        Message Text = [[unixODBC]]
[ODBC][22093][1454628423.118685][SQLError.c][430]
                Entry:
                        Connection = 0x13a5080
                        SQLState = 0x7fff9a5bd7b0
                        Native = 0x7fff9a5bd5a8
                        Message Text = 0x7fff9a5bd5b0
                        Buffer Length = 500
                        Text Len Ptr = 0x7fff9a5bd5ae
[ODBC][22093][1454628423.118704][SQLError.c][467]
                Exit:[SQL_NO_DATA]
[ODBC][22093][1454628423.118722][SQLError.c][510]
                Entry:
                        Environment = 0x13a4750
                        SQLState = 0x7fff9a5bd7b0
                        Native = 0x7fff9a5bd5a8
                        Message Text = 0x7fff9a5bd5b0
                        Buffer Length = 500
                        Text Len Ptr = 0x7fff9a5bd5ae
[ODBC][22093][1454628423.118739][SQLError.c][547]
                Exit:[SQL_NO_DATA]
[ODBC][22093][1454628423.118765][SQLFreeHandle.c][279]
                Entry:
                        Handle Type = 2
                        Input Handle = 0x13a5080
[ODBC][22093][1454628423.118784][SQLFreeHandle.c][330]
                Exit:[SQL_SUCCESS]
[ODBC][22093][1454628423.118827][SQLFreeHandle.c][212]
                Entry:
                        Handle Type = 1
                        Input Handle = 0x13a4750

It successfully allocates a handle, then it tries to connect to the database, fails with a generic SQL_ERROR, tries to do something with the error (not sure what?), and then deallocates the handle.

My last resort was to check the network traffic directly. Here's the initial test, using telnet:

[root@host /opt/ibm/iaccess]# tshark -i tun0 -x
Running as user "root" and group "root". This could be dangerous.
Capturing on tun0
0.000000000   10.10.1.10 -> 123.45.67.8  TCP 60 42054 > ddm-rdb [SYN] Seq=0 Win=13660 Len=0 MSS=1366 SACK_PERM=1 TSval=1992917110 TSecr=0 WS=128

...PACKETS...

2.316931937   10.10.1.10 -> 123.45.67.8  TCP 60 42054 > ddm-rdb [PSH, ACK] Seq=1 Ack=1 Win=13696 Len=8 TSval=1992919427 TSecr=4147650000

0000  45 10 00 3c f1 b3 40 00 40 06 73 d2 0a 0a 01 0a   E..<..@.@.s.....
0010  ac 10 1e 02 a4 46 01 be e0 f5 71 c8 1d a8 3b 71   .....F....q...;q
0020  80 18 00 6b f5 9b 00 00 01 01 08 0a 76 c9 89 83   ...k........v...
0030  f7 38 1d d0 66 6f 6f 62 61 72 0d 0a               .8..foobar..

...PACKETS...

We see some TCP packets, one of which contains foobar, just as we expected.

Now, here's the test with isql:

[root@host /opt/ibm/iaccess]# tshark -i tun0 -x
Running as user "root" and group "root". This could be dangerous.
Capturing on tun0
^C0 packets captured

No traffic!?


Honestly, I'm a bit stuck here. Not sure what to try next. Any thoughts on what could be going wrong, or how to troubleshoot this difficult situation?

Note that the server setup itself is fine. I'm able to connect without a problem using IBM's iSeries ODBC drivers for Windows.

1 Answer 1

0

Alright, finally figured it out. What a doozy.

First of all, do yourself a favor and grab one of IBM's older iSeries drivers, not the newer iAccess ones. When you log in, go to IBM Software > Downloads > No-charge products, tools, and toolkits, then search odbc. You should see drivers like IBM i Access for Linux (V7R1). Grab one of those.

Now, with these older drivers, you get a proper error message:

[08S01][unixODBC][IBM][System i Access ODBC Driver]Communication link failure. comm rc=10060 - CWBCO1048 - A firewall blockage or time-out occurred trying to connect to the IBM i

Nice! At least now we can be absolutely, positively, 120% sure that the issue is some sort of blockage.

But what IS the blockage? IBM's wonderfully well-organized index of documentation to the rescue: http://www-01.ibm.com/support/docview.wss?uid=nas8N1012436

Table from article reproduced here:

PC Function                   Port (non-SSL) SSL Port
Server Mapper                 449            449
License Management (see Note) 8470           9470
RPC/DPC (Remote Command)      8475           9475
Sign-On Verification          8476           9476
Database Access               8471           9471

We unblocked port 8471 ("Database Access") and voila! Everything started working!

Update 3/8/2016: For some reason, Windows needs more ports opened up than Linux. To get this working on Windows, we also needed the Server Mapper and Sign-On Verification ports.


Side note, here's another important paragraph from the article:

These port numbers (other than the Server Mapper) can be configured and, while the defaults are listed above, the actual values on a system may vary. The ports are retrieved from the service table. Determine these ports using the WRKSRVTBLE command on the system in question to determine if the ports have been modified from the default values.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.