"Rene Berber" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]... > "vkd" <[EMAIL PROTECTED]> wrote in message news:<r7zC8.145556$[EMAIL PROTECTED]>... > > Hi, I am experiencing problems using ssh as a shell for CVS. > > > > Everything works great except two things > > > > 1. There is a small hang during connection start up: > > > > # ssh -v -v myuser@remotebox > > OpenSSH_2.9p2, SSH protocols 1.5/2.0, OpenSSL 0x0090602f > > debug1: Reading configuration data /etc/ssh/ssh_config > > debug1: Applying options for * > > debug1: Seeding random number generator > > debug1: Rhosts Authentication disabled, originating port will not be > > trusted. > > debug1: restore_uid > > debug1: ssh_connect: getuid 0 geteuid 0 anon 1 > > > > ---> small hang here <--- about 7-10 seconds > > > > debug1: Connecting to zorro [xxx.xxx.xxx.xxx] port 22. > > debug1: temporarily_use_uid: 0/0 (e=0) > > debug1: restore_uid > > > > Remote box is running Solaris 7 with OpenSSH 3.1p1 > > This is quite normal since (from the OpenSSL documentation) Solaris > doesn't have a /dev/random, generating keys takes a long time. It's > even worse with old Sparc hardware: everytime you login using ssh > (which is what CVS is doing) you can see the CPU usage go to 100%. > > BTW the documentation that comes with OpenSSL (and also de Sun Web > site) recomends installing a /dev/random from one or two sources. It > does help a little.
Sorry I didn't mention. I do have a /dev/random device intalled. It's for a third party, but I am sure you know it sinsce it's the only popular one out there. I can prolly find the URL if you want. It could be this though, since I don't experience the same delay when I go to a linuxbox. I don't get why though. This sun box does have /dev/random. Maybe openssl is not using it? > > 2. CVS transfer completes, but the session just hangs there as though there > > is no completion of communication: > > > > # cvs login > ^^^^^^^^^ > What's this doing here? You only need to log-in if you're using > pserver. Yeah, sorry. That shouldn't be there. I must've gotten it from above. I was testing by using ssh as a tunnel and using pserver and that worked without any hang in the end. > > # cvs co project > > ... > > ... > > U project/utils/readme.txt > > U project/utils/room.c > > U project/utils/zrea.c <--- Last file in project. > > > > ---> indefinate hang here. <--- > > > > What's weird is that if I go into utils area and do cvs update there, it > > works fine. Cvs update results in the same hang when I do it on the whole > > tree. I don't understand why this is going on. > > > > I tried using pserver over ssh tunnel and that worked fine. Why isn't this > > working? > > > > Here is what I have defined: > > > > CVS_RSH=ssh > > CVSROOT=myuser@remotebox:/var/cvsroot > > > > Any ideas? > > > Try by setting CVSROOT=":ext:myuser@remotebox:/var/cvsroot" and then > just do the "cvs co project". > > I don't know why the sample you show hangs, but it has nothing to do > with ssh, it probably is something wrong with cvs. Better test with > the :ext: protocol and see if it still hangs. Well, that's what I did with ssh. CVS_RSH is set to ssh. I am not sure if that :ext: is needed. Now that I updated 2.9 to 3.1, issue seems to be gone. _______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
