[Csync2] Write-error while sending data

[Giampaolo Tomassoni]

Dears,

I have a couple of machines loosely connected through an IPSEC channel (racoon) by a couple of ADSL lines.

These machines actually run linux-2.6.16 from Gentoo.

I'm using csync2-1.32 to keep the web contents of these two servers in sync and, being my channel well authenticated and encripted by IPSEC, I'm using csync in 'clean' channel mode.

That said, I sometimes experience a 'Write-error while sending data'.

I see there was a previous report about this error from Vic Berdin, who started a thread about it on Oct. 25, 2005. That thread seems not have came to find the cause of this trouble, so I tried to dig it my own.

My findings are that this error is not triggered by a crash of the csync2 peer (how suggested at first by Clifford Wolf in the afore-mentioned thread), instead it is due to a SIGALRM signal interrupting a blocking write on the connection channel.

-- Begin excerpt from debug logs --
Local> \252\252\252\252\002\252\252\252\250\n\252\252\252\240*\252\341\325Of3\\235\037\022B\201\017\253\370\205\355\036\367\314\262\r\207\026\324\335t\025U~cf\262\367\256\327\323\353:C\002d\006\207R\361\021O6\215\243\322\313\230\355\307\036\321\217\326\347\352<\243!5}\264h\033f\000\rK'#2,\022\244\260\\311\221\240\320\303E0\026\364F)\224\231\265,\333\0334\221\020\307\340\354\0052\001u\214IM\250B7\200\344\324\023T\314\300\016\265N\243\242\311>\337K\321\000\353\023\346x\204\214\235q\232.\223#G\225\220\271\021\345}\211\364\303\0074K\305\236\244\036\346\343|\203C\270\361r\006\335aN\252f\307\003P1\242\010\262x>\014\242\351\3138\334-\3461\267\2670\260\363\323\036\244\256\205\3434]\341\220\300\330<<\321\345\273x\345\210\305\323\017U\223\247:\037\223\271\213\234\240\013dA\357a\374G\036AR\024\316^\217\017Sds\342?W\304\022v\307\236X\315\202\372\2757\343\021\231\251\215\276\327\332\301\324\220.&\303\361\236\220\0018\363\344\351\217\224\376N\221\226\212\352?\016\226!q\363\007\207\207\326\227\342\023\224\255\337\323\305\326\216(\375\357\337C\255\007\221OTd$,?\037\322\3761\031\212\311\241\361}|9\273\304\265X\313\251\341+\337\324~\033<Z\307\314\036\016\037i^,}gi=p\234rpm\334\231\201UV\225U\020UUUU\001UUUT\005UUUP\025UC\022\016\214\220\200\252\253\3158\274\223\213\350\3126\34184\313B\252\257\235(\261f<=\222\306\343,kRb\205U\}Y\016Psx\206\3166\0167<\021\245\231d*\252\347\276%\231H:\0302q\333\237l\332\316\354\025UE\212s%\277L \342\014\366\331}\344*\251\267\236S%\306d\275\207\014|\021\350G
Database idle in transaction. Forcing COMMIT.
SQL: COMMIT TRANSACTION
Write-error while sending data.
-- End excerpt from debug logs --


The expiring alarm seems somehow related to committing an active SQL transaction, and it seems to happen more or less 2 secs after the file transfer begins. This means that if you have a fast connection to your peer you may probably never see this error since either your files takes less than 2 seconds to tranfer or the write() enters and leaves the blocking state fast enought to keep low the chance of being interrupted.

You'll find a small patch attached to this post which seems to fix the problem. It is made with respect to csync2-1.33.

Note, however, that I'm using a clean connection, so I didn't test it with an ssl one nor I'm that much used to coding with SSL_ calls: I don't know if my patch may even have any meaning in case of SSL channels. So, please, take this patch just as a reference an have a check at it.

Also note that I don't have an idea of the complete csync2 dataflow, so that I can't say whether or not the signal may interrupt even read()s in the code. It seems to work to me, but probably the best thing to do is rework read() and write() code a bit to avoid to be interrupted by signals.

Regards,

-----------------------------------
Giampaolo Tomassoni - IT Consultant
Piazza VIII Aprile 1948, 4
I-53044 Chiusi (SI) - Italy
Ph: +39-0578-21100
-------------- next part --------------
A non-text attachment was scrubbed...
Name: csync2-1.33+write-error-fix.patch
Type: application/octet-stream
Size: 473 bytes
Desc: not available
Url : http://lists.linbit.com/pipermail/csync2/attachments/20061006/21fe9cd3/csync2-1.33write-error-fix.obj