# | Change | User | Description | Committed | |
---|---|---|---|---|---|
#10 | 25550 | C. Thomas Tyler |
Split HMS from SDP, such that HMS will live on as a separate product, layered on the SDP. See: https://swarm.workshop.perforce.com/jobs/SDP-356 Many of the files deleted here have already been copied to HMS. See these HMS changes: https://swarm.workshop.perforce.com/changes/25531 https://swarm.workshop.perforce.com/changes/25533 And this branch spec: SDP_Split_2019.2_HMS In mkrep.sh, HMS-awareness is removed from mkrep.sh, a core SDP script, and the '-f' (Full Setup) flag has been temporarily removed. This useful functionality will be restored in an HMS script. This is a big change, so shellcheck v0.6.0 was implemented to support it. |
||
#9 | 24224 | C. Thomas Tyler |
Enhanced HMS failover to support failover of edge server. Bypassing pre-commit review as this has been tested in the Battle School alb and contains a needed fix. #review-24225 |
||
#8 | 23747 | C. Thomas Tyler |
Fixed hms status issues. Fixed issue getting status of SSL-enabled servers requiring SDP environment (including P4TRUST) to be set, incorporating new irun utility. Fixed issue with illegal negative exit codes in get_component_status(). Added in-code docs for component status values. Fixed get_component_list() to remove call-stack info in high-verbosity mode, due to text output being interpreted as data. |
||
#7 | 23454 | cgeen |
Update to hms script to do two new functions: pull and df. Pull issues a pull -lj on all the edge/replica servers to get the status df issues a diskspace on all the servers. The reason for this is with a hardened edge environment the P4USER does not have permissions to run these commands on the edge servers. Therefore we need to run them centrally as a super. This replaces the replica_status commands on the edge. The updates also includes an update to the p4_vars to change the P4USER on an edge server. The idea is that P4USER only has super privs on the master on the replica edge servers the default user can only have admin privs. This is so the box adminstrator in restricted regions can't change the protection table. Hence the need for pull df centrally as these are the only commands that can't be run on the replicas/edges from the maintance scripts. It also goes without saying that ssh should only go out from the commit and never back. |
||
#6 | 21720 | C. Thomas Tyler |
Fixed bug where a sophsticated topology with mulitple failover targets defined (e.g. 'ha' and 'dr' options) could failover to the wrong host, using the component backup host rather than the one associated with the specified failover target. Bypassing pre-commit review as this has been tested in the Battle School Lab, and I need to push this fix out. #review-21721 |
||
#5 | 21410 | C. Thomas Tyler | Tweaked to support SSL usage. | ||
#4 | 21033 | C. Thomas Tyler |
Fixed hms hang issue during local failover by replacing raw ssh call with a call to the robust runRemoteCmd() function. Also enhanced to detect ServerID based on configured server spec name rather hard-coded value of 'master' used during development. |
||
#3 | 20853 | C. Thomas Tyler | Fixed hms bug discovered in failback testing; master host was wrong. | ||
#2 | 20804 | C. Thomas Tyler | Fixed ssh hang issues. | ||
#1 | 20745 | C. Thomas Tyler |
Approving as is since it isn't changing core SDP functionality, and reviewing it all line by line will take some time. We can do that as we move forward with it. First addition of HMS v1.0 files. This change is a soft launch HMS for initial deployment and testing. Updates to HMS-related files are expected and will bypass pre-commit code review until stabilized. |