Hi, the chain we are investigating here is xrootd->local disks. Othe comments follow: Fabrizio Furano wrote: > Hi all, > > from what I know there are no magics that we can do for this. > Kernel/tcp tuning are just wastes of time. IMHO, one of the main > troubles sources is that the whole chain > > xrootd->gpfs->xfs?->disks > > looks performant when it comes to huge bulk data xfer, but it can > drain everything to crap when stressed with a high transaction rate. > In general, you will never be able to get the same performance which > you were used to when you had the local disks. > Armando, have you tried to monitor how many requests per sec are > being executed? > Looking at xrdlog and netstat output I can roughly see ~10 requests (login and disc) per second > This was an issue, the main imho. Then my suggestions are: > > Put "xrootd.async off" in the data servers. Depending on imper > scrutable factors the async disk read in the OS can give some problems > under stress (see up!), and God only knows what happens with gpfs. > > The value you have for the requesttimeout is a very old default. The > "new" default now (more than 1 year) is 600. But I recall to you that, > under the same setup, to perform our (in)famous performance tests at > CNAF we put 1200. This means that a lot of requests had to wait up to > 20 min. Do you remember what a mess? > > So, my suggestion is 600. This will decrease the "recover" rate. But, > again, you cannot expect magics. Still the clients will recover, > probably less. But then you will try to increase the job rate to the > historical one and will encounter the same problem.... Ok I added in KanAccess.conf: rootenv XNet.RequestTimeout 600 and in /opt/xrootd/etc/xrootd.cf: xrootd.async off and restart the xrootd daemon. > > One more thing: the latest servers contain a fix related to desperate > kernel (?) situations, dealing with chunk sending. If you find (in a > period of at least some hours) some complaints in the xrdlog dealing > with "reset" or something else suspicious, than probably it's a good > idea to upgrade the servers. For the records, it happened on proof > under high stress, and was spotted by Gerri. no suspicious messages about reset, just two of: 080526 13:49:11 3356 XrdProtocol: ?:1789@node132 terminated handshake not received and one of: 080526 13:56:16 3356 XrdPoll: Disabled event occured for babarsgm.11367:5301@node149 The xrootd release: [root@babarxrd ~]# rpm -qa | grep xro xrootd-20071101-0808p1.slac [root@babarxrd ~]# > > One more: the process you are debugging contains many threads, which > spend 99.999% of their life in polling. The interesting thread is just > the main one, not the pollers. In gdb you can recognize the main one > because it executes ROOT calls. All the other ones are insignificant. I'm asking to remote site manager (Pisa Italy) to send me a more complete strace > > Last one: Andy and I put very recently a bugfix in the client for a > rare but nasty problem related to recovering. That was causing a mess > for glast. Wilko, has that problem been solved, even partially? I did > not hear anything yet. (so I can check if you arrived reading up > here.... test and report the magic lottery winning number 123456789! > :-D ) > > Well, but I do not believe that you can upgrade ROOT or recompile it > with an external xrootd package. Let me know. Anyway, if you need it, > just look in Savannah.cern.ch.... > I restarted the daemaon with setting xrootd.async off and client side still 180 seconds of time out, the actual number if job is 144 and the connection are: [root@babarxrd ~]# netstat -natup | grep xro|wc 7726 54082 842134 [root@babarxrd ~]# the following is a tipical log message in a second: 080526 13:56:16 3356 XrootdXeq: babarsgm.5578:6835@node31 login 080526 13:56:16 3356 XrootdXeq: babarsgm.10409:6836@node114 login 080526 13:56:16 3356 XrootdXeq: babarsgm.20472:6837@node142 login 080526 13:56:16 3356 XrootdXeq: babarsgm.4028:6838@gridwn86 login 080526 13:56:16 3356 XrootdXeq: babarsgm.4028:4143@gridwn86 disc 0:01:09 (ended by babarsgm.4028:6838@gridwn86) 080526 13:56:16 3356 XrootdXeq: babarsgm.20472:6613@node142 disc 0:00:57 (ended by babarsgm.20472:6837@node142) 080526 13:56:17 3356 XrootdXeq: babarsgm.29861:4143@node141 login 080526 13:56:17 3356 XrootdXeq: babarsgm.9600:6613@node53 login 080526 13:56:17 3356 XrootdXeq: babarsgm.29861:6760@node141 disc 0:00:43 (ended by babarsgm.29861:4143@node141) 080526 13:56:17 3356 XrootdXeq: babarsgm.9600:6789@node53 disc 0:00:11 (ended by babarsgm.9600:6613@node53) 080526 13:56:17 3356 XrootdXeq: babarsgm.2092:6760@node66 login 080526 13:56:17 3356 XrootdXeq: babarsgm.2092:5999@node66 disc 0:00:17 (ended by babarsgm.2092:6760@node66) 080526 13:56:17 3356 XrootdXeq: babarsgm.12796:5999@node133 login 080526 13:56:17 3356 XrootdXeq: babarsgm.21691:6789@gridwn36 login 080526 13:56:17 3356 XrootdXeq: babarsgm.21691:4243@gridwn36 disc 0:00:20 (ended by babarsgm.21691:6789@gridwn36) 080526 13:56:17 3356 XrootdXeq: babarsgm.28937:4243@node35 login 080526 13:56:17 3356 XrootdXeq: babarsgm.28937:3895@node35 disc 0:00:20 (ended by babarsgm.28937:4243@node35) 080526 13:56:17 3356 XrootdXeq: babarsgm.6690:6839@node71 login 080526 13:56:17 3356 XrootdXeq: babarsgm.6690:6784@node71 disc 0:00:12 (ended by babarsgm.6690:6839@node71) 080526 13:56:18 3356 XrootdXeq: babarsgm.29861:3895@node141 login Thanks to every body, in a day I'll submit new job runs with timeout setting at 600 and I'll post the xrdlog behaviour. In the meanwhile the scenario is still critical. Cheers, Armando > Good luck! > Fabrizio > > > > > >> Date: Sat, 24 May 2008 20:03:53 +0200 >> From: Armando Fella <[log in to unmask]> >> To: Wilko Kroeger <[log in to unmask]> >> Cc: Andrew Hanushevsky <[log in to unmask]> >> Subject: Re: Xrootd problem in SP production >> >> Hi, >> >> the following is part af strace of a SP job asking file to xrootd, it >> hangs in polling. Hope this info cuold help in debugging the problem: >> >> >> connect(14, {sa_family=AF_INET, sin_port=htons(1094), >> sin_addr=inet_addr("212.189.152.199")}, 16) = -1 EINPROGRESS (Operation >> now in progress) >> poll([{fd=14, events=POLLOUT|POLLWRNORM, revents=POLLOUT|POLLWRNORM}], >> 1, 60000) = 1 >> getsockopt(14, SOL_SOCKET, SO_ERROR, [17179869184], [4]) = 0 >> fcntl64(14, F_SETFD, 0x2 /* FD_??? */) = 0 >> time(NULL) = 1210667903 >> time(NULL) = 1210667903 >> poll([{fd=14, events=POLLOUT|POLLERR|POLLHUP|POLLNVAL, >> revents=POLLOUT}], 1, 1000) = 1 >> send(14, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\4\0\0\7\334", 20, 0) = 20 >> time(NULL) = 1210667903 >> time(NULL) = 1210667903 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = 0 >> poll([{fd=14, events=POLLIN}], 1, 1000) = >> ... >> ... >> >> A fresh today xrdlog piece: >> >> 080524 19:29:49 30669 XrootdXeq: babarsgm.30980:780@theown2 disc >> 0:00:04 (ended by babarsgm.30980:756@theown2) >> 080524 19:29:50 30669 XrootdXeq: babarsgm.30615:780@theown5 login >> 080524 19:29:50 30669 XrootdXeq: babarsgm.30615:737@theown5 disc >> 0:00:05 (ended by babarsgm.30615:780@theown5) >> 080524 19:29:52 30669 XrootdXeq: babarsgm.22625:737@theown16 login >> 080524 19:29:52 30669 XrootdXeq: babarsgm.22625:741@theown16 disc >> 0:00:05 (ended by babarsgm.22625:737@theown16) >> 080524 19:29:55 30669 XrootdXeq: babarsgm.10828:741@node54 login >> 080524 19:29:55 30669 XrootdXeq: babarsgm.10828:718@node54 disc >> 0:00:50 (ended by babarsgm.10828:741@node54) >> 080524 19:29:56 30669 XrootdXeq: babarsgm.32501:718@node8 login >> 080524 19:29:56 30669 XrootdXeq: babarsgm.32501:109@node8 disc >> 0:00:32 (ended by babarsgm.32501:718@node8) >> 080524 19:29:58 30669 XrootdXeq: babarsgm.10828:109@node54 login >> 080524 19:29:59 30669 XrootdXeq: babarsgm.32501:781@node8 login >> 080524 19:29:59 30669 XrootdXeq: babarsgm.32501:718@node8 disc >> 0:00:03 (ended by babarsgm.32501:781@node8) >> >> >> >> Cheers, Armando >> >> >> Armando Fella ha scritto: >>> Hi, >>> >>> I'm using the xrootd client enbedded in MooseApp SP 24.2.1l >>> executable. The following is the KanAccess.conf: >>> >>> rootenv Root.XTNetFileAllowWanConnect 1 >>> rootenv Root.XTNetFileAllowWanRedirect 1 >>> rootenv XNet.RedirDomainAllowRE *.infn.it >>> rootenv XNet.ConnectDomainAllowRE *.infn.it >>> rootenv XNet.RequestTimeout 180 >>> >>> read /store/cdb/* xrootd $XROOTD_HOST:1094/ >>> write /store/cdb/* error >>> >>> read /store/cfg/* xrootd $XROOTD_HOST:1094/ >>> write /store/cfg/* error >>> >>> read /store/* xrootd $XROOTD_HOST:1094/ >>> >>> I'm using root version 5.14-00e (but I'm not sure it is the babar >>> one, I should check) >>> The 20 sec is just a case, the login-disc inter time is between 2 >>> min and 3 sec. >>> >>> I tried to use directly the xrd in /opt/xrootd/bin/ on server and on >>> a client wn and it seems to work fine, it transfer the file straight >>> on. >>> >>> I have 2 xrootd servers with this problem and the user bbrmgr (the >>> xrootd launcher) owns the .rootrc file on first server but on the >>> second the file is absent >>> >>> [bbr-serv08] ~ > cat .rootrc >>> Root.XTNetFileAllowWanConnect: 1 >>> Root.XTNetFileAllowWanRedirect:1 >>> [bbr-serv08] ~ > XrootD version: >>> >>> [bbrmgr@babarxrd ~]$ rpm -qa | grep xroo >>> xrootd-20071101-0808p1.slac >>> [bbrmgr@babarxrd ~]$ The system.rootrc is in attachment. >>> >>> The problem is still there and the fail rate is 95% when we reach >>> the limit of 70 jobs asking for cdb cfg bkg on the same xrootd server. >>> >>> Please try to find the bug, we are quite stopped by this problem >>> Please ask if you need more info. >>> >>> We did the following actions: >>> >>> reinstall xrootd >>> reboot the machines >>> modifiy the KanAccess.cfg >>> modify the /etc/sysctl.conf >>> >>> >>> Thanks, Armando >>> >>> Wilko Kroeger ha scritto: >>>> >>>> Hello Armando >>>> >>>> I cc'ed Andy. >>>> >>>> It is not clear what is going wrong. It looks like that it might be >>>> related to the client. >>>> Below are more comments. >>>> >>>> >>>> On Thu, 22 May 2008, Armando Fella wrote: >>>> >>>>> Hi, >>>>> >>>>> I'd add same information: >>>>> >>>>> 1) I try to change the tcp rmem wmem in /etc/sysctl.conf adding >>>>> the lines: >>>>> >>>>> [root@babarxrd ~]# sysctl -p >>>>> net.ipv4.ip_forward = 0 >>>>> net.ipv4.conf.default.rp_filter = 1 >>>>> net.ipv4.conf.default.accept_source_route = 0 >>>>> kernel.sysrq = 0 >>>>> kernel.core_uses_pid = 1 >>>>> net.core.rmem_max = 16777216 >>>>> net.core.wmem_max = 16777216 >>>>> net.ipv4.tcp_rmem = 4096 87380 16777216 >>>>> net.ipv4.tcp_wmem = 4096 65536 16777216 >>>>> net.core.netdev_max_backlog = 250000 >>>>> [root@babarxrd ~]# >>>> >>>> I don't have any experience with tuning the network parameter for >>>> linux so I can't comment on the changes. At slac we use xrootd very >>>> little on linux but the machine we are using it (ER processing, >>>> with ~500 clients) we didn't have to do any tuning). >>>> >>>>> but restarting xrootd the wrong behaviour is still present: >>>>> >>>>> 080522 16:31:01 30669 XrootdXeq: babarsgm.4308:17@theown15 login >>>>> 080522 16:31:01 30669 XrootdXeq: babarsgm.4308:15@theown15 disc >>>>> 0:00:20 (ended by babarsgm.4308:17@theown15) >>>>> 080522 16:31:03 30669 XrootdXeq: babarsgm.9392:15@node75 login >>>>> 080522 16:31:03 30669 XrootdXeq: babarsgm.9392:16@node75 disc >>>>> 0:00:20 (ended by babarsgm.9392:15@node75) >>>>> 080522 16:31:04 30669 XrootdXeq: babarsgm.24721:16@node119 login >>>>> 080522 16:31:04 30669 XrootdXeq: babarsgm.24721:18@node119 disc >>>>> 0:00:20 (ended by babarsgm.24721:16@node119) >>>>> 080522 16:31:17 30669 XrootdXeq: babarsgm.5109:18@gridwn99 login >>>>> 080522 16:31:17 30669 XrootdXeq: babarsgm.5109:19@gridwn99 disc >>>>> 0:00:20 (ended by babarsgm.5109:18@gridwn99) >>>>> 080522 16:31:21 30669 XrootdXeq: babarsgm.4308:19@theown15 login >>>>> 080522 16:31:21 30669 XrootdXeq: babarsgm.4308:17@theown15 disc >>>>> 0:00:20 (ended by babarsgm.4308:19@theown15) >>>>> 080522 16:31:23 30669 XrootdXeq: babarsgm.9392:17@node75 login >>>>> 080522 16:31:23 30669 XrootdXeq: babarsgm.9392:15@node75 disc >>>>> 0:00:20 (ended by babarsgm.9392:17@node75) >>>>> 080522 16:31:24 30669 XrootdXeq: babarsgm.24721:15@node119 login >>>>> 080522 16:31:24 30669 XrootdXeq: babarsgm.24721:16@node119 disc >>>>> 0:00:20 (ended by babarsgm.24721:15@node119) >>>> >>>> It is very suspicious that all the disconnects are after 20 sec. Is >>>> this happening for all SP jobs or only for some of them? >>>> >>>> You could try to copy a file from xrootd with xrdcp and check if >>>> that works fine or if you see a similar behavior. >>>> >>>> Do you know if anything changed on the client site, different root >>>> version, change of timeouts in ~/.rootrc or >>>> $ROOTSYS/root/etc/rootd/system.rootrc >>>> >>>> I guess you are using the babar root version 5.14-00e. Which xrootd >>>> version are you using? >>>> >>>> Sorry that I don't have a more definite answer. >>>> >>>> Cheers, >>>> Wilko >>>> >>>>> >>>>> Cheers, Armando >>>>> >>>>> >>>>> >>>>> Wilko Kroeger ha scritto: >>>>>> >>>>>> Hello Armando >>>>>> >>>>>> Sorry for the late reply. I didn't manage to talk to Andy about >>>>>> it but I will try on Wednesday. >>>>>> I think all your config files and Start scripts look fine. One >>>>>> thing you could do is to remove the >>>>>> all.role server >>>>>> option. By default xrootd starts up as a server and if this >>>>>> directive is omitted the xrootd will not try to connect to an >>>>>> olbd ( >>>>>> 080516 13:17:43 24822 odc_Open: Unable to connect socket to >>>>>> /tmp/.olb/olbd.admin; connection refused) >>>>>> >>>>>> To make sure that your xrootd got started properly you can just >>>>>> do "ps -ef | grep xrootd" and you should only see the -l >>>>>> <logfile> and -c <xrootd.cf> options being used. >>>>>> >>>>>> What I don't understand is from your xrdlog file: >>>>>> >>>>>> 080513 14:36:00 13753 XrootdXeq: babarsgm.15816:3890@gridwn75 login >>>>>> 080513 14:36:00 13753 XrootdXeq: babarsgm.15816:4155@gridwn75 disc >>>>>> 0:00:40 (ended by babarsgm.15816:3890@gridwn75) >>>>>> 080513 14:36:03 13753 XrootdXeq: babarsgm.15816:4155@gridwn75 login >>>>>> 080513 14:36:03 13753 XrootdXeq: babarsgm.15816:3890@gridwn75 disc >>>>>> 0:00:03 (ended by babarsgm.15816:4155@gridwn75) >>>>>> >>>>>> The client babarsgm.15816 was connected but some timeout happened >>>>>> that caused the client to reconnect at 14:36:00 causing its >>>>>> previous session >>>>>> (babarsgm.15816:4155) to be closed. The strange thing now is that >>>>>> after three seconds at 14:36:03 the client reconnects again which >>>>>> causes the previous session (from 14:36:00) to be closed. Three >>>>>> seconds is very short and I don't think that the client or server >>>>>> has any timeout this short. >>>>>> Therefore I also think that >>>>>> rootenv XNet.RequestTimeout 180 >>>>>> would not help. >>>>>> >>>>>> I will talk with Andy maybe he has an idea. >>>>>> >>>>>> Cheers, >>>>>> Wilko >>>>>> >>>>>> >>>>>> On Tue, 20 May 2008, Armando Fella wrote: >>>>>> >>>>>>> Reminder >>>>>>> >>>>>>> Armando Fella ha scritto: >>>>>>>> Here you can find the files I referred in the email: >>>>>>>> >>>>>>>> http://www.cnaf.infn.it/~afella/StartXRD.cf >>>>>>>> http://www.cnaf.infn.it/~afella/StartXRD >>>>>>>> >>>>>>>> Cheers, Armando >>>>>>>> >>>>>>>> Armando Fella ha scritto: >>>>>>>>> *** Discussion title: Simulation Production >>>>>>>>> Email replies to [log in to unmask] must include: >>>>>>>>> In-Reply-To: <[log in to unmask]> >>>>>>>>> Subject: ...change this to be about your reply. >>>>>>>>> >>>>>>>>> This is a multi-part message in MIME format. >>>>>>>>> --------------010909020508030509010608 >>>>>>>>> Content-Type: text/plain; charset=ISO-8859-15; format=flowed >>>>>>>>> Content-Transfer-Encoding: 7bit >>>>>>>>> >>>>>>>>> Hi Wilko, >>>>>>>>> >>>>>>>>> the machine has not been reinstalled or kernel upgraded, the >>>>>>>>> following are the info about OS and architecture: >>>>>>>>> >>>>>>>>> [root@babarxrd ~]# uname -a >>>>>>>>> Linux babarxrd.pi.infn.it 2.6.9-55.EL #1 Thu May 3 23:04:51 >>>>>>>>> CDT 2007 i686 athlon i386 GNU/Linux >>>>>>>>> [root@babarxrd ~]# cat /etc/redhat-release >>>>>>>>> Scientific Linux SL release 4.5 (Beryllium) >>>>>>>>> [root@babarxrd ~]# >>>>>>>>> [root@babarxrd ~]# rpm -qa --queryformat >>>>>>>>> '[("%{NAME}","%{VERSION}-%{RELEASE}","%{ARCH}")\n]' | grep xroo >>>>>>>>> ("xrootd","20071101-0808p1.slac","i386") >>>>>>>>> [root@babarxrd ~]# >>>>>>>>> >>>>>>>>> I suspect that the problem could be in correctness of >>>>>>>>> StartXRD.cf and StartXRD that I got from the xrootd tgz >>>>>>>>> package (can you check that in attachment?), I installed the >>>>>>>>> rpm and I got the two files from the tgz. >>>>>>>>> >>>>>>>>> Is it possible to increase the client timeout to permit the >>>>>>>>> possible network bottleneck to be avoided? >>>>>>>>> The KanAccess.cfg instruction to add is: >>>>>>>>> >>>>>>>>> rootenv XNet.RequestTimeout 180 >>>>>>>>> >>>>>>>>> is 180 the right value? >>>>>>>>> >>>>>>>>> Now I'm not able to increase the site job load so the xrdlog >>>>>>>>> and pstack probably are not so meaningfull, in any case that >>>>>>>>> follow : >>>>>>>>> >>>>>>>>> [root@babarxrd ~]# pidof xrootd >>>>>>>>> 24822 >>>>>>>>> [root@babarxrd ~]# pstack 24822 >>>>>>>>> Thread 16 (Thread -1208767584 (LWP 24823)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b8feac in pthread_cond_timedwait@@GLIBC_2.3.2 () >>>>>>>>> #2 0x08095948 in XrdSysCondVar::Wait () >>>>>>>>> #3 0x0807b04a in XrdBuffManager::Reshape () >>>>>>>>> #4 0x0807a8c0 in XrdReshaper () >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 15 (Thread -1209558112 (LWP 24824)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b8feac in pthread_cond_timedwait@@GLIBC_2.3.2 () >>>>>>>>> #2 0x08095948 in XrdSysCondVar::Wait () >>>>>>>>> #3 0x0808477f in XrdScheduler::TimeSched () >>>>>>>>> #4 0x0808303e in XrdStartTSched () >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 14 (Thread -1210348640 (LWP 24825)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b91b4f in [log in to unmask] () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x08079b60 in XrdSysSemaphore::Wait () >>>>>>>>> #3 0x08083d1a in XrdScheduler::Run () >>>>>>>>> #4 0x08083070 in XrdStartWorking () >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 13 (Thread -1211139168 (LWP 24826)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b91b4f in [log in to unmask] () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x08079b60 in XrdSysSemaphore::Wait () >>>>>>>>> #3 0x08083d1a in XrdScheduler::Run () >>>>>>>>> #4 0x08083070 in XrdStartWorking () >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 12 (Thread -1212376160 (LWP 24827)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00895e04 in poll () from /lib/tls/libc.so.6 >>>>>>>>> #2 0x0808130e in XrdPollPoll::Start () >>>>>>>>> #3 0x0807fb6c in XrdStartPolling () >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 11 (Thread -1213346912 (LWP 24828)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00895e04 in poll () from /lib/tls/libc.so.6 >>>>>>>>> #2 0x0808130e in XrdPollPoll::Start () >>>>>>>>> #3 0x0807fb6c in XrdStartPolling () >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 10 (Thread -1214317664 (LWP 24829)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00895e04 in poll () from /lib/tls/libc.so.6 >>>>>>>>> #2 0x0808130e in XrdPollPoll::Start () >>>>>>>>> #3 0x0807fb6c in XrdStartPolling () >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 9 (Thread -1215108192 (LWP 24830)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b92db6 in __nanosleep_nocancel () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x080963a8 in XrdSysTimer::Wait () >>>>>>>>> #3 0x00210c7d in XrdOdcFinderTRG::Hookup () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x00210a7c in XrdOdcFinderTRG::Start () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #5 0x00210846 in XrdOdcStartOlb () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #6 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #7 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #8 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 8 (Thread -1215898720 (LWP 24831)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b925cb in __read_nocancel () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x08093259 in XrdOucStream::GetLine () >>>>>>>>> #3 0x001f1d85 in XrdOfsEvr::recvEvents () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x001f16d8 in XrdOfsEvRecv () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 7 (Thread -1216689248 (LWP 24832)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b91b4f in [log in to unmask] () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x08079b60 in XrdSysSemaphore::Wait () >>>>>>>>> #3 0x001f1ac3 in XrdOfsEvr::flushEvents () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x001f170a in XrdOfsEvFlush () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 6 (Thread -1217479776 (LWP 24833)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x007ffb30 in do_sigwaitinfo () from /lib/tls/libc.so.6 >>>>>>>>> #2 0x007ffc0f in sigwaitinfo () from /lib/tls/libc.so.6 >>>>>>>>> #3 0x0020c8f2 in XrdOssAioWait () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 5 (Thread -1218270304 (LWP 24834)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x007ffb30 in do_sigwaitinfo () from /lib/tls/libc.so.6 >>>>>>>>> #2 0x007ffc0f in sigwaitinfo () from /lib/tls/libc.so.6 >>>>>>>>> #3 0x0020c8f2 in XrdOssAioWait () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 4 (Thread -1219060832 (LWP 24835)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b92db6 in __nanosleep_nocancel () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x0020dcbe in XrdOssSys::CacheScan () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #3 0x001ff4ec in XrdOssCacheScan () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #4 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #5 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #6 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 3 (Thread -1219851360 (LWP 24836)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b92db6 in __nanosleep_nocancel () from >>>>>>>>> /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x001eba6f in XrdOfsIdleScan () from >>>>>>>>> /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> #3 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #4 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #5 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 2 (Thread -1220641888 (LWP 24837)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b927d8 in accept () from /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x0808d002 in XrdNetSocket::Accept () >>>>>>>>> #3 0x0805c833 in XrdXrootdAdmin::Start () >>>>>>>>> #4 0x0805c4bb in XrdXrootdInitAdmin () >>>>>>>>> #5 0x08095856 in XrdSysThread_Xeq () >>>>>>>>> #6 0x00b8d3cc in start_thread () from /lib/tls/libpthread.so.0 >>>>>>>>> #7 0x0089fc3e in clone () from /lib/tls/libc.so.6 >>>>>>>>> Thread 1 (Thread -1208764736 (LWP 24822)): >>>>>>>>> #0 0x007bd7a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 >>>>>>>>> #1 0x00b927d8 in accept () from /lib/tls/libpthread.so.0 >>>>>>>>> #2 0x0808a1fe in XrdNet::do_Accept_TCP () >>>>>>>>> #3 0x080899de in XrdNet::Accept () >>>>>>>>> #4 0x080893c8 in XrdInet::Accept () >>>>>>>>> #5 0x0807f359 in mainAccept () >>>>>>>>> #6 0x0807f6d1 in main () >>>>>>>>> [root@babarxrd ~]# The xrdlog contains the start daemon >>>>>>>>> messages, no jobs are running now: >>>>>>>>> >>>>>>>>> 080516 13:16:33 001 Scalla is starting. . . >>>>>>>>> Copr. 2007 Stanford University, xrd version 20071101-0808p1_dbg >>>>>>>>> Config using configuration file /opt/xrootd/etc/xrootd.cf >>>>>>>>> ++++++ xrootd [log in to unmask] initialization started. >>>>>>>>> Config maximum number of connections restricted to 65535 >>>>>>>>> Copr. 2007 Stanford University, xrootd version 2.9.0 build >>>>>>>>> 20071101-0808p1 >>>>>>>>> ++++++ xrootd protocol initialization started. >>>>>>>>> =====> xrootd.fslib /opt/xrootd/lib/libXrdOfs.so >>>>>>>>> =====> xrootd.export /store >>>>>>>>> Config warning: 'xrootd.seclib' not specified; strong >>>>>>>>> authentication disabled! >>>>>>>>> Copr. 2007 Stanford University, Ofs Version 20071101-0808p1_dbg >>>>>>>>> ++++++ File system initialization started. >>>>>>>>> =====> all.role server >>>>>>>>> ++++++ Configuring server role. . . >>>>>>>>> Config effective /opt/xrootd/etc/xrootd.cf ofs configuration: >>>>>>>>> ofs.role server >>>>>>>>> ofs.fdscan 9 120 1200 >>>>>>>>> ofs.maxdelay 60 >>>>>>>>> ofs.trace 0 >>>>>>>>> ------ File system server initialization completed. >>>>>>>>> Copr. 2007, Stanford University, oss Version 20071101-0808p1_dbg >>>>>>>>> ++++++ Storage system initialization started. >>>>>>>>> =====> oss.localroot /data >>>>>>>>> =====> oss.path /store r/o >>>>>>>>> 080516 13:16:33 24822 odc_Open: Unable to connect socket to >>>>>>>>> /tmp/.olb/olbd.admin; connection refused >>>>>>>>> Config effective /opt/xrootd/etc/xrootd.cf oss configuration: >>>>>>>>> oss.alloc 0 0 0 >>>>>>>>> oss.cachescan 600 >>>>>>>>> oss.compdetect * >>>>>>>>> oss.fdlimit 32767 65535 >>>>>>>>> oss.maxdbsize 0 >>>>>>>>> oss.localroot /data >>>>>>>>> oss.trace 0 >>>>>>>>> oss.xfr 1 9437184 30 10800 >>>>>>>>> oss.memfile off max 1062334464 >>>>>>>>> oss.defaults r/w nocheck nodread nomig norcreate nostage >>>>>>>>> oss.path /store r/o nocheck nodread nomig norcreate >>>>>>>>> nostage >>>>>>>>> ------ Storage system initialization completed. >>>>>>>>> Config warning: 'xrootd.prepare logdir' not specified; prepare >>>>>>>>> tracking disabled. >>>>>>>>> Config exporting /store >>>>>>>>> ------ xrootd protocol initialization completed. >>>>>>>>> ------ xrootd [log in to unmask]:1094 initialization >>>>>>>>> completed. >>>>>>>>> 080516 13:17:43 24822 odc_Open: Unable to connect socket to >>>>>>>>> /tmp/.olb/olbd.admin; connection refused >>>>>>>>> 080516 13:18:53 24822 odc_Open: Unable to connect socket to >>>>>>>>> /tmp/.olb/olbd.admin; connection refused >>>>>>>>> .... >>>>>>>>> .... >>>>>>>>> .... >>>>>>>>> >>>>>>>>> Thank you for all the help >>>>>>>>> >>>>>>>>> Cheers, Armando >>>>>>>>> >>>>>>>>> Wilko Kroeger ha scritto: >>>>>>>>> >>>>>>>>>> *** Discussion title: Simulation Production >>>>>>>>>> Email replies to [log in to unmask] must include: >>>>>>>>>> In-Reply-To: >>>>>>>>>> <[log in to unmask]> >>>>>>>>>> Subject: ...change this to be about your reply. >>>>>>>>>> >>>>>>>>>> Hello Armando >>>>>>>>>> >>>>>>>>>> The config certainly looks fine and also the load that you >>>>>>>>>> describe should not cause problems. It is a little bit >>>>>>>>>> strange that you suddenly see this problem. Do you know if >>>>>>>>>> anything happened to the machine itself >>>>>>>>>> (update kernel,..)? >>>>>>>>>> >>>>>>>>>> The disconnect messages you see are due to timeouts. The >>>>>>>>>> client does not receive the response from the server within a >>>>>>>>>> certain timeout and therefore it will reconnect to the server >>>>>>>>>> and close (disconnect) its previous session. >>>>>>>>>> >>>>>>>>>> If you still have the problem you could take a pstack >>>>>>>>>> pstack <xrootd_pid> > outfile >>>>>>>>>> and send it (or put some where at slac) me and Andy pstack >>>>>>>>>> output and maybe also the xrdlog file. >>>>>>>>>> I assume the server is a linux machine. >>>>>>>>>> >>>>>>>>>> Cheers, >>>>>>>>>> Wilko >>>>>>>>>> >>>>>>>>>> On Thu, 15 May 2008, Armando Fella wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> we are experiencing unexpected problems on xrootd system >>>>>>>>>>> dedicated to cdb cfg and bkg access in SP production >>>>>>>>>>> (24.2.1l). Three days ago the xrootd server (no oldb) of the >>>>>>>>>>> unique machine serving around 100 SP jobs switched off and >>>>>>>>>>> since its restart the fail rate was around 95%. The server >>>>>>>>>>> worked properly for 6 months with peak 200 jobs asking files. >>>>>>>>>>> >>>>>>>>>>> I checked the parameters in xrootd.cf, but they are in the >>>>>>>>>>> standards: >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# cat /opt/xrootd/etc/xrootd.cf >>>>>>>>>>> xrootd.fslib /opt/xrootd/lib/libXrdOfs.so >>>>>>>>>>> xrootd.export /store >>>>>>>>>>> >>>>>>>>>>> oss.localroot /data >>>>>>>>>>> oss.path /store r/o >>>>>>>>>>> >>>>>>>>>>> all.role server >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# The xrdlog shows continuously login and >>>>>>>>>>> disconnecting connection: >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# tail xrdlog >>>>>>>>>>> 080513 14:36:00 13753 XrootdXeq: babarsgm.3924:4170@gridwn24 >>>>>>>>>>> login >>>>>>>>>>> 080513 14:36:00 13753 XrootdXeq: babarsgm.3924:3890@gridwn24 >>>>>>>>>>> disc 0:01:49 (ended by babarsgm.3924:4170@gridwn24) >>>>>>>>>>> 080513 14:36:00 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15816:3890@gridwn75 login >>>>>>>>>>> 080513 14:36:00 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15816:4155@gridwn75 disc 0:00:40 (ended by >>>>>>>>>>> babarsgm.15816:3890@gridwn75) >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15816:4155@gridwn75 login >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15816:3890@gridwn75 disc 0:00:03 (ended by >>>>>>>>>>> babarsgm.15816:4155@gridwn75) >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15208:3890@gridwn65 login >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: >>>>>>>>>>> babarsgm.15208:4141@gridwn65 disc 0:00:36 (ended by >>>>>>>>>>> babarsgm.15208:3890@gridwn65) >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: babarsgm.3924:4141@gridwn24 >>>>>>>>>>> login >>>>>>>>>>> 080513 14:36:03 13753 XrootdXeq: babarsgm.3924:4170@gridwn24 >>>>>>>>>>> disc 0:00:03 (ended by babarsgm.3924:4141@gridwn24) >>>>>>>>>>> [root@babarxrd ~]# >>>>>>>>>>> The xrootd version is: >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# rpm -qa | grep xroot >>>>>>>>>>> xrootd-20071101-0808p1.slac >>>>>>>>>>> [root@babarxrd ~]# >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Other symptoms are the quick raising of connection (70 jobs): >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# netstat -natup | grep xroot |wc >>>>>>>>>>> 3715 26005 404935 >>>>>>>>>>> [root@babarxrd ~]# lsof -i | grep xroot |grep >>>>>>>>>>> gridwn131.pi.infn.it|wc >>>>>>>>>>> 104 936 12480 >>>>>>>>>>> [root@babarxrd ~]# lsof -i | grep xroot |grep >>>>>>>>>>> gridwn84.pi.infn.it|wc >>>>>>>>>>> 205 1845 24395 >>>>>>>>>>> [root@babarxrd ~]# lsof -i | grep xroot |grep >>>>>>>>>>> gridwn76.pi.infn.it|wc >>>>>>>>>>> 115 1035 13685 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> dstat output (not so much throughput): >>>>>>>>>>> >>>>>>>>>>> [root@babarxrd ~]# ./dstat >>>>>>>>>>> ----total-cpu-usage---- -dsk/total- -net/total- ---paging-- >>>>>>>>>>> ---system-- >>>>>>>>>>> usr sys idl wai hiq siq| read writ| recv send| in out | >>>>>>>>>>> int csw >>>>>>>>>>> 40 1 58 2 0 0| 25k 27k| 0 0 | 0 >>>>>>>>>>> 3.7B|1699 265 >>>>>>>>>>> 0 4 96 0 0 0| 0 64k| 31k 520k| 0 0 >>>>>>>>>>> |1937 391 >>>>>>>>>>> 0 5 95 0 0 0| 0 16k| 38k 619k| 0 0 >>>>>>>>>>> |2109 480 >>>>>>>>>>> 0 4 96 0 0 0| 16k 0 | 34k 638k| 0 0 >>>>>>>>>>> |2045 349 >>>>>>>>>>> 0 3 96 1 0 0| 24k 120k| 36k 592k| 0 0 >>>>>>>>>>> |2068 419 >>>>>>>>>>> 1 3 96 0 0 0| 0 248k| 36k 547k| 0 0 >>>>>>>>>>> |2068 482 >>>>>>>>>>> 0 3 96 1 0 0| 20k 24k| 33k 580k| 0 0 >>>>>>>>>>> |2001 344 >>>>>>>>>>> 0 4 96 0 0 0|4096B 8192B| 39k 660k| 0 0 >>>>>>>>>>> |2126 437 >>>>>>>>>>> [root@babarxrd ~]# >>>>>>>>>>> Client side the KanAccess.cfg file is: >>>>>>>>>>> >>>>>>>>>>> rootenv XNet.RedirDomainAllowRE *.infn.it >>>>>>>>>>> rootenv XNet.ConnectDomainAllowRE *.infn.it >>>>>>>>>>> >>>>>>>>>>> read /store/cfg/* xrootd $XROOTD_HOST:1094/ >>>>>>>>>>> write /store/cfg/* error >>>>>>>>>>> >>>>>>>>>>> read /store/cdb/* xrootd $XROOTD_HOST:1094/ >>>>>>>>>>> write /store/cdb/* error >>>>>>>>>>> >>>>>>>>>>> read /store/* xrootd $XROOTD_HOST:1094/ >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> please hints. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Thanks, Armando >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> -- >>>>>>>>>>> =================================================== >>>>>>>>>>> Armando Fella (BaBar support at INFN-CNAF Tier-1) >>>>>>>>>>> --------------------------------------------------- >>>>>>>>>>> Viale Berti Pichat 6/2, 40127, Bologna >>>>>>>>>>> office 3, via Ranzani 13/2 >>>>>>>>>>> >>>>>>>>>>> Email: armando.fella at cnaf.infn.it >>>>>>>>>>> armando.fella at pi.infn.it >>>>>>>>>>> armando.fella at gmail.com >>>>>>>>>>> >>>>>>>>>>> Phone in Bologna: +39 051 6092 902 >>>>>>>>>>> Phone in Pisa: +39 050 2214 231 >>>>>>>>>>> =================================================== >>>>>>>>>>> >>>>>>>>>>> --------------------------------------------------------------------- >>>>>>>>>>> Unless unavoidable, no Word, Excel or PowerPoint >>>>>>>>>>> attachments, please. >>>>>>>>>>> See http://www.gnu.org/philosophy/no-word-attachments.html >>>>>>>>>>> >>>>>>>>>>> --------------------------------------------------------------------- >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>>> -- >>>>>>> =================================================== >>>>>>> Armando Fella (BaBar support at INFN-CNAF Tier-1) >>>>>>> --------------------------------------------------- >>>>>>> Viale Berti Pichat 6/2, 40127, Bologna >>>>>>> office 3, via Ranzani 13/2 >>>>>>> >>>>>>> Email: armando.fella at cnaf.infn.it >>>>>>> armando.fella at pi.infn.it >>>>>>> armando.fella at gmail.com >>>>>>> >>>>>>> Phone in Bologna: +39 051 6092 902 >>>>>>> Phone in Pisa: +39 050 2214 231 >>>>>>> =================================================== >>>>>>> >>>>>>> --------------------------------------------------------------------- >>>>>>> >>>>>>> >>>>>>> Unless unavoidable, no Word, Excel or PowerPoint attachments, >>>>>>> please. >>>>>>> See http://www.gnu.org/philosophy/no-word-attachments.html >>>>>>> >>>>>>> --------------------------------------------------------------------- >>>>>>> >>>>>>> >>>>>>> >>>>> >>>>> -- >>>>> =================================================== >>>>> Armando Fella (BaBar support at INFN-CNAF Tier-1) >>>>> --------------------------------------------------- >>>>> Viale Berti Pichat 6/2, 40127, Bologna >>>>> office 3, via Ranzani 13/2 >>>>> >>>>> Email: armando.fella at cnaf.infn.it >>>>> armando.fella at pi.infn.it >>>>> armando.fella at gmail.com >>>>> >>>>> Phone in Bologna: +39 051 6092 902 >>>>> Phone in Pisa: +39 050 2214 231 >>>>> =================================================== >>>>> >>>>> --------------------------------------------------------------------- >>>>> >>>>> Unless unavoidable, no Word, Excel or PowerPoint attachments, please. >>>>> See http://www.gnu.org/philosophy/no-word-attachments.html >>>>> >>>>> --------------------------------------------------------------------- >>>>> >>>>> >>> >> -- ==================================================== Armando Fella (BaBar support at INFN-CNAF Tier-1) ---------------------------------------------------- Viale Berti Pichat 6/2, 40127, Bologna office 3, via Ranzani 13/2 Mail: armando.fella at cnaf.infn.it Phone in Bologna: +39 051 6092 902 Phone in Pisa: +39 050 2214 330 ====================================================