Project

General

Profile

Task #17913

test staging latency feature introduced in 4.2.2 FileRouter

Added by Natalia Ratnikova about 2 years ago. Updated about 2 years ago.

Status:
Assigned
Priority:
High
Assignee:
-
Target version:
-
Start date:
10/16/2017
Due date:
10/23/2017
% Done:

0%

Estimated time:
Spent time:
Duration: 8

Description

The feature has been deployed without operational testing on my side : I tested that the new value has been propagated properly. We need to confirm that it results in the desired behavior, i.e. router will now favor non-MSS sites. and adjust the value as needed.
Work with transfer team and Dave on this in the next week.

History

#3 Updated by Natalia Ratnikova about 2 years ago

Here is a record in the router log when the option was deployed:

-bash-4.1$ grep -C 3 STAGING_LATENCY /data/ProdNodes/Prod_Mgmt/logs/mgmt-router 
2017-10-16 11:31:44: InfrastructureFileRouter[30853]: debug: Agent was loaded DB=>1, CYCLE=1, DROPBOX=1
2017-10-16 11:31:44: FileRouter[30853]: (re)connecting to database
2017-10-16 11:31:44: FileRouter[30853]: Creating new private DBH
2017-10-16 11:32:23: FileRouter[30853]: router configuration: WINDOW_SIZE=50.00 TB, REQUEST_ALLOC=BY_AGE, MIN_REQ_EXPIRE=7.00 hours, EXPIRE_JITTER=3.00 hours, LATENCY_THRESHOLD=3.00 days, STAGING_LATENCY=604800 secs, PROBE_CHANCE=0.02, DEACTIV_ATTEMPTS=5, DEACTIV_TIME=30.00 days, NOMINAL_RATE=0.50 MB/s, N_SLOW_VALIDATE=3 paths
2017-10-16 11:32:50: FileRouter[30853]: executed flush:  0 stuck blocks suspended, 0 unwanted requests deleted, 311 deleted requests with replica, 0 request priority updated, 0 stray requests added, 547 invalid paths validated, 12460 path expire extended, 12460 task expire extended, 12460 request expire extended, 14093 expired requests deactivated, 37661 invalid paths deleted, 0 exclusions deleted
2017-10-16 11:32:50: FileRouter[30853]: warning: Unable to activate 40 blocks for routing to node 71: no active incoming links
2017-10-16 11:32:50: FileRouter[30853]: warning: Unable to activate 1 blocks for routing to node 102: no active incoming links
--
2017-10-16 13:12:40: InfrastructureFileRouter[31166]: debug: Agent was loaded DB=>1, CYCLE=1, DROPBOX=1
2017-10-16 13:12:40: FileRouter[31166]: (re)connecting to database
2017-10-16 13:12:40: FileRouter[31166]: Creating new private DBH
2017-10-16 13:13:11: FileRouter[31166]: router configuration: WINDOW_SIZE=50.00 TB, REQUEST_ALLOC=BY_AGE, MIN_REQ_EXPIRE=7.00 hours, EXPIRE_JITTER=3.00 hours, LATENCY_THRESHOLD=3.00 days, STAGING_LATENCY=604800 secs, PROBE_CHANCE=0.02, DEACTIV_ATTEMPTS=5, DEACTIV_TIME=30.00 days, NOMINAL_RATE=0.50 MB/s, N_SLOW_VALIDATE=3 paths
2017-10-16 13:13:44: FileRouter[31166]: executed flush:  0 stuck blocks suspended, 0 unwanted requests deleted, 1743 deleted requests with replica, 0 request priority updated, 0 stray requests added, 896 invalid paths validated, 7003 path expire extended, 7003 task expire extended, 7003 request expire extended, 1771 expired requests deactivated, 13191 invalid paths deleted, 1358 exclusions deleted
2017-10-16 13:13:44: FileRouter[31166]: warning: Unable to activate 40 blocks for routing to node 71: no active incoming links
2017-10-16 13:13:44: FileRouter[31166]: warning: Unable to activate 1 blocks for routing to node 102: no active incoming links
-bash-4.1$ 

The output of the routerhistory data service is save in

-bash-4.1$ ls -l  /afs/cern.ch/user/r/ratnik/WORK/DATAOPS/staging_fnal_github_1100/routerhistory 
-rw-r--r--. 1 ratnik zh 1370690 Oct 16 15:24 /afs/cern.ch/user/r/ratnik/WORK/DATAOPS/staging_fnal_github_1100/routerhistory



Also available in: Atom PDF