95515 - double entry in Manage CRLs after importing single CRL

Status: RESOLVED WORKSFORME

(NSS :: Libraries, defect, P2)

Description

[awnuk]

I don't have exact steps to reproduce this problem,
but this is quite consistent case.
I was just playing with importing CRLs and delta-CRLs,
deleting them and importing new CRLs and deltas again.
Eventually after importing single CRL I saw in Manage CRLs
two identical entries for this CRL.

Assuming database corruption, I removed all db-files
after each incident, but after several imports
I was getting double entries in Manage CRLs again and again.

If you need running CMS to play with,
you can find it at http://pancake.red.iplanet.com:8400/

Comment 1

[Stephane Saux]

->future

Comment 2

[awnuk]

This might be a serious case database corruption.
It is relatively easy to reproduce so it should not be postponed.  

Comment 3

[Keyser Sose]

marking NEW.

Comment 4

[Stephane Saux]

rangansen

Comment 5

[Rangan Sen]

Actually, right now NSS does not support/handle delta CRLs, though it allows
them to be downloaded [bug# 108031]. To my understanding, when a crl is stored
in the db, the issuer org name is used as the key. My guess is that this might
cause problem when full and delta crls from the same issuer org is imported. 

Reporter: do you see the problem if delta crls are not used - i.e., when the
same base crl is imported and deleted multiple times?

Comment 6

[Rangan Sen]

The related NSS bug is bug# 103946 and not 108031 - sorry about that...

Comment 7

[Rangan Sen]

-> NSS

Comment 8

[Wan-Teh Chang]

Assigned the bug to Bob.  Bob, is this something we
can address in NSS 3.4?

Comment 9

[Robert Relyea]

This is weird, because the Key is the DN of the server. It would require a
pretty serious bug in DB before this can happen. The only way I can see for this
to happen is if the crls somehow have subtly different Subjects.

To answer Wan-Teh's question, The whole Crl management code has been ripped up
for 3.4, so I would consider this part of making sure we don't have any
regressions in the new 3.4 code.

bob

Comment 10

[Robert Relyea]

NSS 3.4 is checked in on the tip, delta-CRL's are now properly rejected, can you
still reproduce this problem?

Comment 11

[Wan-Teh Chang]

Changed the QA contact to Bishakha.

Comment 12

[Robert Relyea]

Give there was not response now that we reject delta CRL's, I'm closing this bug.