Changes between Initial Version and Version 1 of tier-one-meeting-20190326


Ignore:
Timestamp:
03/26/19 16:13:50 (21 months ago)
Author:
paul
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • tier-one-meeting-20190326

    v1 v1  
     1[[TOC(depth=0)]] 
     2= dCache Tier I meeting MONTH DATE, 2013 = 
     3[part of a [wiki:developers-meetings series of meetings]] 
     4 
     5== Present == 
     6 
     7dCache.org(Paul), IN2P3(), Sara(), Triumf(), BNL(), NDGF(Jens), PIC(), KIT(Samuel), Fermi(), CERN() 
     8 
     9= Agenda = 
     10 
     11(see box on the other side) 
     12 
     13= Site reports = 
     14 
     15== NDGF == 
     16 
     17Things are pretty happy. 
     18 
     19=== sweeper purge DoS attack === 
     20 
     21We had an issue yesterday. 
     22 
     23Benchmarking -- checking pool start-up time. 
     24 
     25Many small pools: tried 1 pool, 5 pools, 20 pools 
     26 
     27Yes, this greatly improves start-up times. 
     28 
     29When the test was over -- "sweeper purge" to clear cache. 
     30 
     31Unfortunately, "sweeper purge" lead to severe impact on dCache . 
     32 
     33Central services were affected, complaining about PNFS messaging timeouts (e.g., PinManager complaining). 
     34 
     35After 10 minutes, dCache recovered. 
     36 
     37Last configurable was with 20 pools. 
     38 
     39There were 850,000 files in total  
     40 
     41Also found some stack-traces.  Will report these. 
     42 
     43 
     44== KIT == 
     45 
     46Everything is running fine. 
     47 
     48=== ZK issue === 
     49 
     50Currently running stand-alone zookeeper cluster with three nodes.  We want to move one ZK server to another rack; in effect, decommission one zookeeper server and commission a new ZK server. 
     51 
     52Currently, this requires restarting all dCache domains, since the ZK connection string is changing.  Could we use a DNS aliases (one for each ZK server) to avoid having to restart dCache domains? 
     53 
     54=== 9648 === 
     55 
     56Accessing the file twice with the same UUID. 
     57 
     58Host running test every 15 minutes. 
     59 
     60Noticed changing the xrootd debug level from 2 to 3, now seeing the UUID 
     61 
     62= Support tickets for discussion = 
     63 
     64[Items are added here automagically] 
     65 
     66= DTNM = 
     67 
     68Same time, next week.