Possible DCM4CHEE-ARC-LIGHT BUG Study Size in KB Calculation INCORRECT

[Michael-Andersen]

I have been encountering an issue where Study Size in KB is incorrect on many studies. I know the value is wrong because I have summed the instance level tags and also checked the size on the disk manually.

I have regular automated polling of metadata hitting my PACS instance every few minutes and I have reason to believe this might be causing the issue but there is no log of an error in dcm4chee-arc-light 33.1 to indicate that anything is going wrong. In another instance at version 32.0 I do see some database connection resource errors near study size calculation so it could be that newer versions have suppressed a legitimate error message?

If you need more info I can try to find those database connection errors in the logs and explain further how to reproduce the error.

Any help on this issue would be greatly appreciated, Thanks!

[vrindanayak]

Check both

Study Size in KB (7777,1029) UL
Study Size Bytes (7777,102A) US

• Study Size Bytes (777,102A) does not contain the size of the study in bytes. The size of the study in bytes = Study Size in KB (777,1029) * 1000 + Study Size Bytes (777,102A)
• Study Size in KB (7777,1029) actually contains the result of the integer division study-size-in-bytes / 1000. DICOM UL VR has a value length of 4 bytes, which may be too small to encode the value of the study size in bytes. That's the reason for splitting the value in 2 DICOM attributes.

[Michael-Andersen]

Thanks for the clarification. Unfortunately the additional bytes tag doesn't make up the discrepancies we sometimes see. Study Size in KB (7777,1029) UL is sometimes significantly off, sometimes showing less than 50% of the actual size. Recreating the DB record usually fixes the tag but sometimes has unintended side effects (see my other discussion topic).

[vrindanayak]

Study size calculation can be retriggered directly using UI Navigation->Studies->SIZE or by scheduler by configuring the Calculate Study Size Polling Interval / Calculate Study Size Delay. Looking at the response from other topic as well, I still do not understand why recreate DB record needs to be used.

[Michael-Andersen]

Recreate DB record doesn't need to be used, I just tried to use it as it seemed like a possible solution to this issue, the point of my other topic was too highlight that it appeared buggy to me, regardless of whether it is an appropriate solution to this problem.

How do I retrigger the size calculation directly from the UI ("UI Navigation->Studies->Size" I don't quite follow what you mean, other than viewing the size tags in the UI which doesn't automatically correct them for us)?