[9fans] DNS failures

[9fans] DNS failures

[lucio]

I'm getting consistent failures from
	ndb/dns -rs
that look like this:

	assert failed: rp->magic == RRmagic

The corresponding stack dump is:

	ripple# acid 78867
	/proc/78867/text:386 plan 9 executable
	/sys/lib/acid/port
	/sys/lib/acid/386
	acid: lstk()
	abort()+0x0 /sys/src/libc/9sys/abort.c:6
	_assert(s=0x35f06)+0x3a /sys/src/libc/port/_assert.c:12
	rrattach1(new=0x16dd60,auth=0x1)+0x217 /sys/src/cmd/ndb/dn.c:808
		dp=0xb9510
		l=0x118b28
		rp=0xdc758
	rrattach(rp=0x16dd60,auth=0x1)+0x4e /sys/src/cmd/ndb/dn.c:872
		next=0xdb6c8
		dp=0xb9510
	procansw(mp=0xdfffd9bc,p=0x98e30,qp=0x11b1a8,srcip=0xdfffd9f0,depth=0x1)+0x114 /sys/src/cmd/ndb/dnresolve.c:1236
		soarr=0x0
		ndp=0x0
		tp=0xa26efd8
		buf=0x12a23
		nqp=0x182f1c
		rv=0xd6
	queryns(qp=0x11b1a8,obuf=0xdd908,depth=0x1,inns=0x1,waitms=0x9e4,ibuf=0x182ee8)+0x3b8 /sys/src/cmd/ndb/dnresolve.c:1449
		req=0xda484b18
		len=0x26
		dest=0x98e30
		p=0x98e30
		ndest=0x1
		endms=0xe945ab05
		replywaits=0x0
		buf=0x0
		m=0x4b18
		srcip=0x0
		rv=0xdfffda48
	udpquery(mntpt=0x3f388,qp=0x11b1a8,patient=0x0,depth=0x1,inns=0x1)+0x1b7 /sys/src/cmd/ndb/dnresolve.c:1579
		ibuf=0x182ee8
		obuf=0xdd908
		fd=0x10
		msg=0x17779
		pcntprob=0x5f
		reqtm=0x1f40
		wait=0x9e4
		rv=0xdd388
	netquery(depth=0x1,qp=0x11b1a8)+0x2b5 /sys/src/cmd/ndb/dnresolve.c:1661
		rv=0x0
		dp=0xb9510
		qlp=0xb954c
		lock=0x1
		buf=0x17437
		triedin=0x0
		inname=0x1
	netqueryns(qp=0x11b1a8,nsrp=0x1128a8,depth=0x1)+0x1e /sys/src/cmd/ndb/dnresolve.c:338
		rv=0x99530
	issuequery(class=0x1,qp=0x11b1a8,depth=0x0,name=0xb9650,recurse=0x0)+0x1e9 /sys/src/cmd/ndb/dnresolve.c:413
		nsrp=0x1128a8
		cp=0xb965a
		dbnsrp=0x0
		rp=0x11b1a8
	dnresolve1(name=0xb9650,type=0x1,class=0x1,req=0xdfffdd48,depth=0x0,recurse=0x0)+0x25c /sys/src/cmd/ndb/dnresolve.c:505
		dp=0xb9510
		rp=0x0
		qp=0x11b1a8
	dnresolve(status=0x0,depth=0x0,rooted=0x1,name=0xb9650,class=0x1,type=0x1,req=0xdfffdd48,cn=0xdfffdd7c,recurse=0x0)+0x299 /sys/src/cmd/ndb/dnresolve.c:228
		procname=0xab530
		rp=0xdc7c8
		drp=0x0
		nrp=0x0
		nname=0x0
		dp=0x92810
		loops=0x0
	doextquery(mp=0xdfffdd68,req=0xdfffdd48,recurse=0x0)+0x64 /sys/src/cmd/ndb/dnserver.c:186
		rp=0x0
		neg=0xf46e2
	dnserver(repp=0xdfffdd68,reqp=0xdfffdd90,rcode=0x0,srcip=0xdfffddb8,req=0xdfffdd48)+0x1fa /sys/src/cmd/ndb/dnserver.c:83
		recursionflag=0x80
		tp=0xb0f70
		tname=0x0
		myarea=0x0
		neg=0x0
		cp=0x3fd28
		nsdp=0x0
	dnudpserver(mntpt=0x3f388)+0x56d /sys/src/cmd/ndb/dnudpserver.c:257
		fd=0xf
		req=0x1
		repmsg=0x249
		reqmsg=0x249
		buf=0x0
		len=0x25
		uh=0xdfffddb8
		rcode=0x0
		err=0x0
		op=0x0
		tname=0x0
		p=0x3fd28
	main(argv=0xdfffefb0,argc=0x0)+0x319 /sys/src/cmd/ndb/dns.c:264
		ext=0x0
		_argc=0x73
		_args=0xdfffefc7
		servefile=0x642f7323
		dir=0x0
		kid=0x0
	_main+0x31 /sys/src/libc/386/main9.s:16
	acid:

It is likely that I have a consistency error in the database:
something along these lines, but not very clearly, was reported when I
tried to do a zone transfer from Plan 9 to NetBSD; I have not been
able to track the cause down.

Perhaps somebody here has a clue on how best to approach solving this
problem?

++L

Re: [9fans] DNS failures

[erik quanstrom]

> It is likely that I have a consistency error in the database:
> something along these lines, but not very clearly, was reported when I
> tried to do a zone transfer from Plan 9 to NetBSD; I have not been
> able to track the cause down.

yes, there clearly is junk in the in-memory database.  that line is

		assert(rp->magic == RRmagic && rp->cached);

i would recommend using acid to see which one of these is false.
dumping the whole rp would make sense.

- erik