Problems Generating MySql Database tables

stevenw · Jan 22nd 2009

I followed the install for linux from this link http://www.image-systems.biz/f…it=tutorial+install#p4124
but when I run ./dgate -v -r I get this:
Regen Database
Step 1: Re-intialize SQL Tables
Dropping Existing tables (if-any)
Worklist is empty
Dropping worklist
Dropping other tables
WorkList Database
Patient Database
Study Database
Series Database
Image Database
***DB error
Step 2: Load / Add DICOM Object files
Regen Device 'MAG0'
***[Regen] ./data/HEAD_EXP_00097038/0001_003000_892665662.v2 -FAILED: Error SQL Add
***[Regen] ./data/HEAD_EXP_00097038/0001_002000_892665661.v2 -FAILED: Error SQL Add
Regeneration Complete

The database tables have not been created. I check the username password used different usernames and passwords I also checked to make sure the the database name was correct but nothing seams to give the dgate access to my mysql database. I can asses the database form the command line or from phpmysqladmin page without a problem. Anyone have any ideas?

Steven

marcelvanherk · Jan 22nd 2009

Hi,

just to make sure, can you try another database such as sqlite and see what happens?

Marcel

stevenw · Jan 22nd 2009

Ok I installed sqlite and reinstalled conquest and I get this warning.

warning: warning: fattach is not implemented and will always fail

Is this a problem?

I started sending studies over to conquest and it was working but then it stopped working and I got with error

Code

UPACS THREAD 132: STARTED AT: Thu Jan 22 12:08:17 2009
A-ASSOCIATE-RQ Packet Dump
        Calling Application Title : "NGHNRHP         "
        Called Application Title : "Drapper         "
        Application Context : "1.2.840.10008.3.1.1.1", PDU length: 16384
        Number of Proposed Presentation Contexts: 1
        Presentation Context 0 "1.2.840.10008.5.1.4.1.1.2" 1
Server Command   := 0001
Message ID       := 0001
FreeStore Left 152056 on ./data/
***Truncated AccessionN from 18 to 16 chars in file: 18215/1.2.840.113619.2.30.1.1762369785.1752.1205779002.288_0002_000061_12326548970084.v2
Written file: ./data/18215/1.2.840.113619.2.30.1.1762369785.1752.1205779002.288_0002_000061_12326548970084.v2
UPACS THREAD 132: ENDED AT: Thu Jan 22 12:08:17 2009
UPACS THREAD 132: TOTAL RUNNING TIME: 0 SECONDS
***SQLITEExec error: Ð.´¶mn StudyInsta is not unique
*** glibc detected *** ./dgate: double free or corruption (fasttop): 0xb6b454a0 ***
======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6[0xb7d56a85]
/lib/tls/i686/cmov/libc.so.6(cfree+0x90)[0xb7d5a4f0]
./dgate[0x80e88e2]
./dgate[0x8136d1e]
./dgate[0x8098608]
./dgate[0x80994d3]
./dgate[0x80a272b]
./dgate[0x80a32b7]
./dgate[0x80cc42d]
./dgate[0x80e1775]
./dgate[0x8071b60]
./dgate[0x80cf83e]
./dgate[0x804ebea]
./dgate[0x804ec23]
/lib/tls/i686/cmov/libpthread.so.0[0xb7f664fb]
/lib/tls/i686/cmov/libc.so.6(clone+0x5e)[0xb7dc1e5e]
======= Memory map: ========
08048000-0815b000 r-xp 00000000 09:02 239549     /usr/local/conquest/dgate
0815b000-0815e000 rw-p 00112000 09:02 239549     /usr/local/conquest/dgate
0815e000-085fe000 rw-p 0815e000 00:00 0          [heap]
b5b00000-b5c00000 rw-p b5b00000 00:00 0
b5d00000-b5e00000 rw-p b5d00000 00:00 0
b6100000-b6200000 rw-p b6100000 00:00 0
b62ff000-b6300000 ---p b62ff000 00:00 0
b6300000-b6c00000 rw-p b6300000 00:00 0
b6c9b000-b6c9c000 ---p b6c9b000 00:00 0
b6c9c000-b749c000 rw-p b6c9c000 00:00 0
b749c000-b749d000 ---p b749c000 00:00 0
b749d000-b7ceb000 rw-p b749d000 00:00 0
b7ceb000-b7e34000 r-xp 00000000 09:02 10330136   /lib/tls/i686/cmov/libc-2.7.so
b7e34000-b7e35000 r--p 00149000 09:02 10330136   /lib/tls/i686/cmov/libc-2.7.so
b7e35000-b7e37000 rw-p 0014a000 09:02 10330136   /lib/tls/i686/cmov/libc-2.7.so
b7e37000-b7e3a000 rw-p b7e37000 00:00 0
b7e3a000-b7e44000 r-xp 00000000 09:02 10330129   /lib/libgcc_s.so.1
b7e44000-b7e45000 rw-p 0000a000 09:02 10330129   /lib/libgcc_s.so.1
b7e45000-b7e68000 r-xp 00000000 09:02 10330140   /lib/tls/i686/cmov/libm-2.7.so
b7e68000-b7e6a000 rw-p 00023000 09:02 10330140   /lib/tls/i686/cmov/libm-2.7.so
b7e6a000-b7f52000 r-xp 00000000 09:02 99251      /usr/lib/libstdc++.so.6.0.9
b7f52000-b7f55000 r--p 000e8000 09:02 99251      /usr/lib/libstdc++.so.6.0.9
b7f55000-b7f57000 rw-p 000eb000 09:02 99251      /usr/lib/libstdc++.so.6.0.9
b7f57000-b7f5d000 rw-p b7f57000 00:00 0
b7f5d000-b7f5f000 r-xp 00000000 09:02 10330139   /lib/tls/i686/cmov/libdl-2.7.so
b7f5f000-b7f61000 rw-p 00001000 09:02 10330139   /lib/tls/i686/cmov/libdl-2.7.so
b7f61000-b7f75000 r-xp 00000000 09:02 10330150   /lib/tls/i686/cmov/libpthread-2.7.so
b7f75000-b7f77000 rw-p 00013000 09:02 10330150   /lib/tls/i686/cmov/libpthread-2.7.so
b7f77000-b7f7a000 rw-p b7f77000 00:00 0
b7f7e000-b7f80000 rw-p b7f7e000 00:00 0
b7f80000-b7f81000 r-xp b7f80000 00:00 0          [vdso]
b7f81000-b7f9b000 r-xp 00000000 09:02 10331605   /lib/ld-2.7.so
b7f9b000-b7f9d000 rw-p 00019000 09:02 10331605   /lib/ld-2.7.so
bf817000-bf82c000 rw-p bffeb000 00:00 0          [stack]
Aborted

Display More

Can you tell me what is going on.

stevenw · Jan 23rd 2009

I also checked my message log and found this:

Code

Jan 22 11:17:07 Drapper kernel: [668125.585336] dgate[11640]: segfault at 08048af0 eip b7cd9da4 esp bf940660 error 7
Jan 22 11:23:34 Drapper kernel: [668513.430001] dgate[11675]: segfault at 00000046 eip b7d3d1e9 esp b7490768 error 4

Any ideas?

marcelvanherk · Jan 23rd 2009

Hi,

there are two problems:

1) your data is not correct, this causes: "StudyInsta is not unique" (the same UID occurs in two patients), and "Truncated AccessionN from 18 to 16 chars in file...". These messages are OK. See, e.g., http://www.image-systems.biz/forum/viewtopic.php?f=33&t=618.

2) The fault indicates that the handling code for these data errors (which are very rare), is not OK for SQLite - this code has apperently not been exercised yet - I have added a bug report. Can you reproduce it from gdb and get a "bt" listing?. The mysql error is probably due to the same data error and there it handles it correctly. So, I would suggest going back to mysql or trying dbaseIII as well.

The fattach code is a leftover from the original UC Davis code that is not used, this warning is OK.

Marcel

stevenw · Jan 26th 2009

I can try to reproduce it but how do I do it from gdb and get the "bt" string?

marcelvanherk · Jan 26th 2009

Hi, if I remember correctly:

gdb dgate
run -v
(wait until crash)
bt

Marcel

stevenw · Jan 26th 2009

OK I am currently running gdb on dgate no crash so fare. Also for info when I ran dgate -v -r this was my output:

Code

Regen Database
Step 1: Re-intialize SQL Tables
Dropping Existing tables (if-any)
Worklist is empty
Dropping worklist
Dropping other tables
WorkList Database
Patient Database
Study Database
Series Database
Image Database
Step 2: Load / Add DICOM Object files
Regen Device 'MAG0'
***Truncated AccessionN from 18 to 16 chars in file: 7001/1.2.840.113789.16020.115580365422920077063081204160594623478_0001_000002_1232674622000c.v2
[Regen] ./data/7001/1.2.840.113789.16020.115580365422920077063081204160594623478_0001_000002_1232674622000c.v2 -SUCCESS
Segmentation fault

Display More

I don't know if this might shed some light on the problem

stevenw · Jan 26th 2009

OK on this

Quote

"StudyInsta is not unique" (the same UID occurs in two patients)

is this the patient ID that is sent over to conquest.

I know that from our last Pac system the system made if very difficult to change the Patient ID if it was entered in on the modality wrong, so there are quite a few patients with the wrong ID's. We are in the process of getting them fixed but have not got them all. Will this cause problems with conquest.

marcelvanherk · Jan 26th 2009

Hi,

Awaiting the "bt".

The other error occurs if the same data is stored twice with different patient IDs. This will cause conquest to reject one of the two. This is not a conquest issue but a general dicom issue. Conquest itself can safely modify patient IDs (generating new UIDs).

Marcel

stevenw · Jan 26th 2009

OK here is what gdb gives when dgate crashes:

Code

#0  0xb7f5c410 in __kernel_vsyscall ()
#1  0xb7cf2085 in raise () from /lib/tls/i686/cmov/libc.so.6
#2  0xb7cf3a01 in abort () from /lib/tls/i686/cmov/libc.so.6
#3  0xb7d2ab7c in ?? () from /lib/tls/i686/cmov/libc.so.6
#4  0xb7d32a85 in ?? () from /lib/tls/i686/cmov/libc.so.6
#5  0xb7d364f0 in free () from /lib/tls/i686/cmov/libc.so.6
#6  0x080e88e2 in sqlite3GenericFree ()
#7  0x08136d1e in sqlite3_free ()
#8  0x08098608 in Database::SQLITEExec ()
#9  0x080994d3 in Database::AddRecord ()
#10 0x080a272b in UpdateOrAddToTable ()
#11 0x080a3200 in SaveToDataBase ()
#12 0x080cc42d in SaveToDisk ()
#13 0x080e1775 in MyUnknownStorage::CheckObject ()
#14 0x08071b60 in UnknownStorage::Read ()
#15 0x080cf83e in StorageApp::ServerChild ()
#16 0x0804ebea in DriverApp::ServerChildThread ()
#17 0x0804ec23 in DriverHelper ()
---Type <return> to continue, or q <return> to quit---
#18 0xb7f424fb in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#19 0xb7d9de5e in clone () from /lib/tls/i686/cmov/libc.so.6

Display More

marcelvanherk · Jan 26th 2009

Thanks for the bt,

but now I am going to bed early, I will have a look at it later..

Marcel

marcelvanherk · Feb 3rd 2009

Hi,

I believe the error message was freed twice. Please replace routine SQLITEExec at line 3424 in odbci.cpp the following modified code:

Code

int	Database :: SQLITEExec(char *StatementText, BOOL q)
	{
	int i, rc;
	char *err=NULL;


	if (resultp)
		{ 
		sqlite3_free_table(resultp);
		resultp = NULL;
		}


	if (!(rc=sqlite3_get_table(sqlitedata, StatementText, &resultp, &ntuples, &ncolumns, &err)))
		{
		if (!q && resultp) 
			{ 
			sqlite3_free_table(resultp);
			resultp = NULL;
			ntuples = 0;
			ncolumns = 0;
			current_tuple = 0;
			return 0;
			}
		current_tuple = 0;
		return 0;
		}

	if (rc==SQLITE_BUSY) 
		for(i=0; i<60; i++)
			{
			sqlite3_free(err);
			err = NULL;


			delay(100 + rand_r(&seedp)%2000);
			if (!(rc=sqlite3_get_table(sqlitedata, StatementText, &resultp, &ntuples, &ncolumns, &err)))
				{
				if (!q && resultp) 
					{ 
					sqlite3_free_table(resultp);
					resultp = NULL;
					ntuples = 0;
					ncolumns = 0;
					current_tuple = 0;
					return 0;
					}
				current_tuple = 0;
				return 0;
				}
			}

	OperatorConsole.printf("***SQLITEExec error: %s\n", err);
	sqlite3_free(err);


	return -1;
	}

Display More

And recompile. Hope this solves it.

Marcel

stevenw · Feb 4th 2009

I made the code changes and compiled but when I ran dgate -v -r to regen the database this is what displayed

Code

Regen Database
Step 1: Re-intialize SQL Tables
Dropping Existing tables (if-any)
Worklist is empty
Dropping worklist
Dropping other tables
WorkList Database
Patient Database
Study Database
Series Database
Image Database
Step 2: Load / Add DICOM Object files
Regen Device 'MAG0'
***Truncated AccessionN from 18 to 16 chars in file: 301450/1.2.840.113619.2.5.17162218.1207318607.603.5_0603_000046_123300
[Regen] ./data/301450/1.2.840.113619.2.5.17162218.1207318607.603.5_0603_000046_123300078401b1.v2 -SUCCESS
***Truncated AccessionN from 18 to 16 chars in file: 301450/1.2.840.113619.2.80.45793822.18432.1207344590.1.4.1_0400_000033
Segmentation fault

Display More

When I run dgate -v it seams to run but none of my previous studies are available and there was several hundred that are in there. Also I tried dbase3 and that seamed to work just fine had it running for several days without a crash so I am fine using dbase instead of sqlite.

I do have a couple of questions. I am using this for live archival purposes only. The only time that this server will be used other then to store images is if an old study needs to be retrieved is there any real advantage using sqlite over dbase. The HD size of our current images are about 300 to 400 GB increasing by 5 to 10 GB a month. Also I have 3 pac systems that I will be moving studies over from, a lot of the studies are the same study and there are to many to remove, will there be a problem just sending them all to conquest server.

Steven

marcelvanherk · Feb 4th 2009

Hi,

apparently there are two bugs... Can you get me a gdb bt of this fault as well? Just want to fix everything we run into prior to the next release.

The real advantage of sqlite is that it is a real sql server: it can handle more complex queries, and the database size is more compact. But no, for your application dbase will do just fine, up to a couple of million images. For bigger archives MySQL is preferred.

Sending stuff twice is no problem - the first sent data will just be overwritten.

Marcel

stevenw · Feb 4th 2009

ok

Here is the bt from gdb when I run dgate -v -r

Code

#0  0xb7d55da4 in strcpy () from /lib/tls/i686/cmov/libc.so.6#1  0x08097d48 in Database::NextRecord ()#2  0x080a1984 in UpdateOrAddToTable ()#3  0x080a32bd in SaveToDataBase ()#4  0x080a368f in RegenToDatabase ()#5  0x080a39b8 in RegenDir ()#6  0x080a3977 in RegenDir ()#7  0x080a3bef in Regen ()#8  0x080a408d in Regen ()#9  0x080d77d6 in ParseArgs ()#10 0x080d9a98 in main ()

Also I ran dgate -v through gdb and it stop about an hour into running with this error.

Code

Program received signal SIGINT, Interrupt.

Here is the bt on it

Code

#0  0xb7f08410 in __kernel_vsyscall ()
#1  0xb7ef5bb8 in accept () from /lib/tls/i686/cmov/libpthread.so.0
#2  0x0807f3d6 in Socket::Accept ()
#3  0x0809a8e3 in DriverApp::Server ()
#4  0x080d9e3e in main ()

marcelvanherk · Feb 5th 2009

Thanks,

I will try to puzzle it out!

Marcel

marcelvanherk · Feb 8th 2009

Hi,

I cannot reproduce this error here. Can you send me the image which crashes the system?

301450/1.2.840.113619.2.80.45793822.18432.1207344590.1.4.1_0400_000033

Marcel

Problems Generating MySql Database tables

Participate now!

Share