Hi Burton: As for #4, it also happens to me. I am running ntop in a Red Hat 7.2, kernel 2.4.7-10, on a PII-400 SMP (dual processor). When I start without -j flag, more or less after half an hour ntop crashes. The process remains there, but the http server doesn't answer anymore and I have to kill -SIGKILL the process. The flags I use are: -u ntop -L -d -S 2 -w 3000 -n -m x.x.x.x/x -p /usr/local/ntop/protocol.list. When I use -j flag, it doesn't crash. I have place ntop at our internet link (5 Mbps).
Starting with -K I got the following messages: ntop[22538]: ntop caught signal 11 ntop[22538]: *****ntop error: Signal(11) ntop[22538]: backtrace is: ntop[22538]: 1. /lib/i686/libpthread.so.0 [0x402638d5] ntop[22538]: 2. /lib/i686/libc.so.6 [0x40542848] ntop[22538]: 3. /lib/i686/libc.so.6(__libc_free+0xb4) [0x40594bf4] ntop[22538]: 4. /usr/local/ntop/lib/libntop-2.0.so.99(ntop_safefree+0x51) [0x402b6509] ntop[22538]: 5. /usr/local/ntop/lib/libntop-2.0.so.99(freeHostInfo+0x20d) [0x402b1ea5] ntop[22538]: 6. /usr/local/ntop/lib/libntop-2.0.so.99(purgeIdleHosts+0x2ed) [0x402b26f1] ntop[22538]: 7. /usr/local/ntop/lib/libntop-2.0.so.99(scanIdleLoop+0x65) [0x402b98c5] ntop[22538]: 8. /lib/i686/libpthread.so.0 [0x40260b9c] ntop[22538]: 9. /lib/i686/libc.so.6(__clone+0x3a) [0x406011ea] ntop[22538]: Cleaning up... ntop[22538]: Waiting until threads terminate... ntop[22538]: Freeing hash host instances... (1 device(s) to save) I'd like to run it under gdb, but I am not a gdb guru... Regards, --Inaki. "Burton M. Strauss III" wrote: > > (e.g. stuff to work on pre 2.1) > > 1. SSLWATCHDOG -- (Burton) In progress > Trap the hang-up from NS6.2.2 et al in the ntop web server. > I have a working version but would like to make it use ONE thread, vs. > one per request. > > 2. facilitynames / glibc / __gnuc__ -- (Burton) COMMITTED 2Jul2002 > Change the configure/Makefile to test for the presence of > facilitynames[], vs. just assuming it's available in a __gnuc__ environment. > > 3. Segmentation fault - Luca patched 3Jul2002 - awaiting reports. > > ERROR: Index mismatch (hostTrafficBucket=69/theIdx=820) > and > ERROR: purgeHostIdx(0,820) failed [host not found] > > User (Charles Grijalv) asked to run under gdb, referred to Luca. > > 4. ntop slows down and crashes on large network, 768MB of RAM/Dual Celeron > 466, Linux 2.4.18. > > Seems to crash many times while looking at the web interface. Could be > part of the interface isn't SMP safe under high loads?? > > User (Michael Baird) asked to run under gdb or -K or with backtrace... > pending > > 5. <hostSymIpAddress>Bridge Sp. Tree/OSI Route <IMG SRC="/card.gif" > ALT="Network Int</hostSymIpAddress> in XML dump... Patch sent to user to > check as I can't duplicate > > My fault - I updated the generated stuff so that the html would have a > W3C required ALT tag, and missed the use in the dump... fix sent to user > for validation... > > 6. (Linux only, typically RedHat, one report from Debian) (frequently under > debug, sometimes just plain running) ntop hangs freeing an entry in the > myGlobals.tcpSvc structure (it loops in malloc.c routine chunk_free(arena > *ar_ptr, mchunkptr p), specifically looping at the call to frontlink(ar_ptr, > p, sz, idx, bck, fwd); > > No clue... > > Anyway, as soon as I get #1 fixed, I'm going to commit it and post an RC3... > > -----Burton > > _______________________________________________ > Ntop mailing list > [EMAIL PROTECTED] > http://lists.ntop.org/mailman/listinfo/ntop _______________________________________________ Ntop mailing list [EMAIL PROTECTED] http://lists.ntop.org/mailman/listinfo/ntop
