 |
 |
Hi Les, Kaushik De <[log in to unmask]> is involved with US ATLAS production, so he may be able to suggest a suitable host. regards, Stephen. On Wed, 23 Aug 2006, Cottrell, Les wrote: > Thanks, I found for the US: Boston U, Harvard U, Chicago U, Indiana U, Univ Texas at Arlington, Oklahoma U, Univ of New Mexico, Langston U, and I think SLAC, UCSD and LBNL should be added, plus BNL & CERN. Are there others? > > I can start with those. Maybe later if there is interest I can add other country Tier 1 & 2s. > > I need hosts at those sites that will respond to pings. Typically I use the web server. However pings are often blocked. Looking at the web servers Univ Texas at Arlington (www.uta.edu), Univ of New Mexico (www.unm.edu) and LBL (www.lbl.gov) all block pings, the others are OK. For LBL I can use ns1.lbl.gov, and for Univ Texas at Arlington I can use ns1.uta.edu. So I need someone to give me the name of a host at UNM that is always up and responds to pings. Do you have a contact I can work with? > > Currently we do not monitor any of these sites (we do monitor SDSC which is on the UCSD campus) using PingER (see http://www-iepm.slac.stanford.edu/pinger/), so we will need to add them to the list of hosts to be monitoted and also assign an ATLAS group to them. We could also add making traceroutes to them at 10 min intervals to assist in diagnosing problems/events. The goal if we do this is to enable long term tracking and visualization of simple performance measures between SLAC and the other sites. This can be very valuable for detecting when something changed to see if it correlates with a user perception of degraded performance. We are also working on detecting anomalous events on the end-to-end paths by analyzing the time series for changes. We will be looking for persistent events as opposed to momentary changes in performance due to say congestion. Apart from complete loss of connectivity, due to the low frequency of measurements (at 30 minute intervals in order ! to! > limit > network load) the events (step changes in performance seen in the time series) will be detected several hours after they occur. Emails can be sent to interested paries. Once we have some results (e.g. measurements going back a month or so, then we can add a pointer to the SLAC ATLAS Tier 2 web site. > > There is an ATLAS group already for PingER. It includes: BNL, UCSD, CERN, TRIUMF, ITEP (Russia), RAL, LFN.INFN among others. One way to view the existing results is to go to http://www-iepm.slac.stanford.edu/cgi-wrap/pingtable.pl?file=packet_loss&by=by-node&size=100&tick=monthly&from=WORLD&to=ATLAS&ex=none&dataset=hep&percentage=any One can choose other metrics (RTT, loss etc.) from there plus other time ticks (aggregated hourly, monthly, daily etc.) for the data, monitoring sites and remote sites or groups of sites. > > Is this of interest to the SLAC ATLAS community? It's not a lot of work, but if nobody cares then probably we should not embark on it or put it on some back-burner. From my viewpoint I would like to do it, I believe it will be useful and give PingER more exposure, but will need assistance to answer questions, add links, etc. from the SLAC ATLAS community. > > -----Original Message----- > From: Su, Dong > Sent: Wednesday, August 23, 2006 10:20 AM > To: Cottrell, Les; Yang, Wei; atlas-sccs-planning-l > Subject: RE: Tier 2 web page > > There was a pointer buried in my replies to Stephen's Aug/9 meeting minutes which may be useful for locating the US Tier-2 sites: > http://www.usatlas.bnl.gov/twiki/bin/view/Admins/WebHome#Tier2_Site_Web_Pages > There is a separate page for BNL Tier-1 > http://www.acf.bnl.gov/ > but I am not sure either are really up to date. > Su Dong > > >> -----Original Message----- >> From: [log in to unmask] >> [mailto:[log in to unmask]] On Behalf Of >> Cottrell, Les >> Sent: Wednesday, August 23, 2006 9:58 AM >> To: Yang, Wei; atlas-sccs-planning-l >> Subject: RE: Tier 2 web page >> >> Is there a list of ATLAS Tier 1 and 2 sites such that we could set up >> a web page showing connectivity, round-trip-time, loss, jitter to >> those sites from SLAC, CERN & BNL? >> >> -----Original Message----- >> From: [log in to unmask] >> [mailto:[log in to unmask]] On Behalf Of >> Yang, Wei >> Sent: Tuesday, August 22, 2006 9:16 PM >> To: atlas-sccs-planning-l >> Subject: Tier 2 web page >> >> I discussed with Len about the Tier 2 web page. The issue can be >> divided into two areas: content management tools and Tier >> 2 content. Here is a summary. any comment? >> >> Content management tools: >> >> A static page is good at the beginning. In the near future, we might >> want to look at the possibility of using Plone, which is a Wiki-like >> tool but provides more features. >> >> Tier 2 content: >> >> The discussion focused on the needs of 'local users'. But now I am not >> so sure if this is correct. I will add what I think about the needs of >> grid users at the end. >> >> -------------------------------------------------------------- >> --------- >> Notification >> *) Outage, major changes >> *) Events >> >> How to obtain computer accounts for SLAC Tier 2 >> *) Unix account + e-mail account, prefer to send e-mail to users' home >> institutes. >> *) Full SLAC accounts (unix, e-mail, Windows/exchange, link to >> existing >> page) >> *) Procedures to obtain accounts (discussion: PI -> Charlie -> >> HelpDesk ?) >> >> SLAC Tier 2 Facilities >> *) SLAC computing environment, short text and a link to existing one. >> *) Security page, a link to existing page. >> *) Public machines >> *) Setup Atlas environment, links to Stephen's page and Atlas workbook >> *) Disk space >> in general >> AFS related issues >> Atlas space areas >> *) Batch >> LSF documents >> Commands to submit Atlas jobs to SLAC LSF farm >> LSF resources available to Atlas local users >> >> Data Availability >> *) DQ2 browser and space (Panda monitoring page) >> *) How to bring in and transfer out datasets (discussion: do we allow >> a >> local user to do this?) >> >> Helps >> *) HyperNews at CERN >> *) [log in to unmask] >> *) Other contact info >> >> Userful Links >> ... >> -------------------------------------------------------------- >> ----------- >> Grid users should not need much info about us. I am thinking to start >> with the following: >> >> *) Outage, major change, events. >> *) How to obtain a certificate/account for grid jobs to SLAC >> including limitations of SLAC grid accounts. >> *) Data availability (see above for local users) >> *) Submit jobs via Panda, a link to BNL >> *) HyperNews at CERN for discussion >> *) BNL RT (for Western T2) for Tier 2 related help >> *) Panda page/DQ2 page for various statistics. >> *) Ganglia monitoring in the future? >> >> -- >> Wei Yang | [log in to unmask] | 650-926-3338(O) >> >> >> >> > -- /------------------------------------+-------------------------\ |Stephen J. Gowdy, SLAC | CERN Office: 32-2-A22| |http://www.slac.stanford.edu/~gowdy/ | CH-1211 Geneva 23 | |http://calendar.yahoo.com/gowdy | Switzerland | |EMail: [log in to unmask] | Tel: +41 22 767 5840 | \------------------------------------+-------------------------/