1 / 9

TGC COOL field to folder mapping

TGC COOL field to folder mapping. Daniel Lellouch , David Front 5 Oct 2011. Context: Considering how to write all TGC DCS conditions to COOL. TBD: How map fields to folders ?. Write all TGC DCS conditions To COOL. DCS to PVSS archive. dcs. PVSS a rchive DB. COOL DB.

barry-witt
Télécharger la présentation

TGC COOL field to folder mapping

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. TGC COOL field to folder mapping Daniel Lellouch, David Front 5 Oct 2011

  2. Context: Considering how to write all TGC DCS conditions to COOL • TBD: How map fields to folders ? Write all TGC DCS conditions To COOL DCS to PVSS archive dcs PVSS archive DB COOL DB • TBD: Can the 15 minutes delay be reduced to a few minutes? Current situation: DCS to TGC DB TBD: Can TGC API read DCS conditions directly from COOL, or will COOL conditions be written to TGC DB? COOL to TGC DB TGC DB Online Offline 2

  3. Average daily amount of TGC conditions during September 2011 • 2 * means that there are 2 underlying dpe types: • The same type of HW is read in two different ways, depending on its geometrical location • TGC DB does not have this differentiation, it hides such HW complexities.Data is kept in a more logically simple manner.All channelids of any field may be queried easily together in one query. Hence, for TGC, moving to use DCS data from COOL may complicate the way it will retrieve data. 3

  4. field to folder mapping 1 : Per field folders • Each field has its folder: ~30folders • + Simple to read from • Show stopper: • Fields from different DPE types may not reside at same folder • Question: What is the reason for this restriction? • Number of folders should not be > 10 per sub detector • Question: What is the reason for this restriction? • - Resources not optimally utilized: • each condition of a field has a record at COOL 4

  5. field to folder mapping 2: Per field per DPE type folders each field/PDE type has a one field folder: ~60 folders + Simple to read from - Considerably more folders than 10 - Resources not optimally utilized: each condition of a field has a record at COOL 5

  6. field to folder mapping 3: Per DPE type folders • Each DPE type has its folder: ~15 folders • + Minimizes number of folders • Mixing fields into a folder that are not written at same time causes extra IOVS with extra data to keepand may cause queries to be more heavy thanfor single field folders 6

  7. field to folder mapping 4: Separate frequently inserted fields • Each DPE type has its folder • and fields that insert frequently • are separated into other folders: ~35 folders • + Less ‘extra IOVs’ • Not trivial mapping of fields to folders, complicates reading • (Same type of HW maps to two different DPE types) • Number of folders considered to be too high 7

  8. A summary of the options 8

  9. Summary • For TGC, writing DCS conditions to COOL has the benefits of reusing the same data path that other sub detectors use, and avoiding the TGC specific SW for writing from PVSS to TGC DB • However, it complicates the (t.b.d.)SW which reads from COOL: • Same type of HW maps to two different DPE types • Mapping of fields to folders has to be considered for each query • We may see the need to write the COOL conditions to TGC DB, instead of a direct read • We would like to understand better why fields from different DPE types may not reside at same folder: without this restriction, the number of folders could be reduced by about one half and this would solve issue 2. a. • We would like to understand better the motivation for restricting the number of folders per sub detector to <= 10. Without this restriction, we could have ‘DB efficient’ folders • We prefer option 3 ‘Per field per PDE type folders’, if possible; otherwise option 4 ‘Per PDE type with added folders for fields that insert frequently’ . 9

More Related