[Halld] Fwd: Log entry notification: HALLD Web and DB Server Status

[Mark Ito]

Folks,

The new server is halldweb.jlab.org, not halldweb1.jlab.org. The old 
name is supposed to auto-switch you to the new one. I thought that 
worked this morning but does not seem to be working as I write this. The 
new wiki link is thus:

  https://halldweb.jlab.org/wiki/index.php/Main_Page

On the wiki, you now log in with your JLab CUE username and password. 
Really should have told you guys that before now. Our wiki is now 
standard in that regard.

   -- Mark


On 02/25/2015 07:36 AM, Mark Ito wrote:
> We should be up now...let us know if you find anomalies.
>
> -------- Original Message --------
> Subject: Log entry notification: HALLD Web and DB Server Status
> From: wise at jlab.org
> To: unixadm at jlab.org,hovanes at jlab.org,marki at jlab.org,davidl at jlab.org
> CC:
>
>
>
> *1022640 
> <http://cclog.jlab.org//elog.php?quicksearch=1022640&logbook=ALL&entry_type=ALL#bracketed>* 
> 	*HALLD Web and DB Server Status*
> *25-Feb-15 06:52* 	*M_Wise* 	*SYSTEMLOG | MAINTLOG*
>
>
>
> The deployment of the new web and DB servers for Hall D is nearly
> complete with the following exceptions --
>
> * The midas electronic logbook is not configured on halldweb yet
> * GLUONWEB still remain to be set up. Apache is installed, working on the
> MariaDB installation now.
> * Replication of the CCDB database to GLUONDB1 is not operational yet
>
> Otherwise, the new servers are up and running. This includes HALLDWEB,
> HALLDDB and the externally-visible replication slave HALLDDB-EXT (note:
> off-site, this is called HALLDDB). The new systems are all running new
> versions of everything and have improvements in available resources (more
> CPU, RAM, etc. appropriate to each) and improvements in the backup
> systems, replication configuration, etc.
>
> The main CC-maintained applications are all up and working, including the
> Gluex Wiki (https://halldweb.jlab.org/wiki), the HDOPS wiki (now at
> https://halldweb.jlab.org/hdops/wiki), and MantisBT
> (https://halldweb.jlab.org/mantisbt). All are running the newest
> versions. I've gone through several of the other applications
> (JInventory, SP_Jinventory, etc) and updated their DB connection
> parameters so they are now working as well.
>
> The hallddb-ext and halldweb systems are both visible from off-site and
> have had all the necessary DNS entries updated.
>
> The original systems (halldweb1, hdops and hallddb1) have all been
> preserved, but access to them from user VLANs has been blocked using
> their local firewalls. Also, halldweb1's DNS entry was changed to
> halldweb1-old.jlab.org. This was done so that the original name could be
> aliased to the new system. So halldweb1.jlab.org is an alias for
> halldweb.jlab.org.
>
> *** NOTES:
>
> 1. DBs all originally were on halldweb1. I split them up between halldweb
> and hallddb. The intention is that only DBs that are used exclusively by
> applications running on halldweb will be maintained on the local DB
> server on halldweb. Other DBs should move to hallddb for general access.
> It may be that I guessed incorrectly about some of these, so we may need
> to shuffle a few around yet.
>
> 2. DB user accounts were migrated from the old system. A few, obviously
> obsolete accounts were removed. Also, many needed to be updated to
> reflect the new client system names. This is a tedious, manual process.
> Access should be reviewed and any problems reported so I can correct them.
>
> 3. DB accounts with no passwords, or granted to all hosts (e.g.
> someuser@% should be carefully reviewed. Since the hallddb system is
> replicated to the slave entirely, grants to all hosts now apply to
> internet clients. The server is set to read-only in its config, but,
> nevertheless, such accounts should be removed or changed to a limited
> scope (like %.jlab.org) unless specifically required to support off-site
> access.
>
> 4. All DB accounts from halldweb1 were copied to both halldweb and
> hallddb. Many can be removed form the system where they are not used.
>
> I will get gluonweb configured, then replication to gluondb1 and finally
> the midas elog. I will email as these additional components are brought
> online.
>
> Let me know of any problems seen. Things seem to be working well
> fundamentally, but I'm sure there are small details that will need
> correction and/or adjustment.
>
>
> *References to other log entries:*
> *1022636 
> <elog.php?quicksearch=1022636&logbook=ALL&entry_type=ALL#bracketed>* 
> *HALLD Web and DB Server Status*
> *24-Feb-15 21:39* 	*M_Wise* 	*SYSTEMLOG | SCI-COMPLOG | MAINTLOG*
>
>
>
> HALLDWEB, HALLDDB and HALLDDB-EXT are all online. All still need some
> checkout and finishing up, but the 2 DB servers are operational and
> replicating to the ext server. FW is opened and DNS updated, but can't
> connect yet... still getting old IP on my home system.
>
> The Web Server is up and running, and the two wikis -- Gluon wiki and
> the HDOPS wiki are both working to first order, though I need to do some
> more checkout.
>
> Other apps need edits to their configs to reflect the new system name
> (halldweb). I will work through the remainder of them in the morning.
>
> Replication of the wiki DB to gluonweb is in progress. I have the apache
> server installed and preliminary configuration done on gluon24 but need
> to install the DB server.
>
> When I get the remaining items finished up, I'll send an email with the
> new URLs. I will also install redirects or announcement pages on the old
> URLs.
>
>
> *Other log entries that reference this one:*
> 1022640 - HALLD Web and DB Server Status <elog_item.php?elog_id=1022640>
>
>
>
>
> Database query + PHP page generation time = 0 seconds
>
>
> _______________________________________________
> Halld mailing list
> Halld at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/halld/attachments/20150225/95aad7d7/attachment.html