Uploaded image for project: 'Confluence Data Center'
  1. Confluence Data Center
  2. CONFSERVER-8637

Resin 3 responses are being concatenated sometimes when a custom logo is specified

      The keep-alive option seems to be affected when a large images are used.

      For some reason the byte-size of the actual image is causing failure. When a smaller image is used, 8K png works just fine. Please investigate.

            [CONFSERVER-8637] Resin 3 responses are being concatenated sometimes when a custom logo is specified

            Don Willis added a comment -

            The fix for CONF-11325 should fix this bug too.

            Don Willis added a comment - The fix for CONF-11325 should fix this bug too.

            Don Willis added a comment -

            This behaviour is almost certainly caused by CONF-11325, which we're prioritising very highly.

            Don Willis added a comment - This behaviour is almost certainly caused by CONF-11325 , which we're prioritising very highly.

            We've ran out of time we can expend on the issue and have begun the migration back onto Tomcat 5.5 unless there are specific requests to stick on Resin ... although there haven't been any so far. In future we hope that we can re-evaluate this decision and get back onto Resin.

            Dan Hardiker added a comment - We've ran out of time we can expend on the issue and have begun the migration back onto Tomcat 5.5 unless there are specific requests to stick on Resin ... although there haven't been any so far. In future we hope that we can re-evaluate this decision and get back onto Resin.

            Pam Cole added a comment -

            We are still encountering this error several times an hour, even after upgrading to Resin 3.1.4:

            Receive error messages similar to the following after clicking Edit or Add New Page:

            Ézn°Ý¸)[Þ„|ÿ3ý8uÝA1©LÜØÚ''vC½.u‡-‡K|P8 ä ЯJºÒU_t- }ˆ ïÁ}FÙÐRùQ¦ž]§¡wÈ%ÀÑØXqú Sµ`€³Qˆ...çågP&ôR‹ô! & žˆ™(zMD7­=ø ÌG£&6:ÖšCV¡Ö"ÿï©PÛŒCFW^irUåÔ'X¶Ä$ +0bO`†)&˜ X'÷Pp&! 0Â;ÔY'vÚé" ÙfXZùЛq.2'2ˆ&ÊPðÉ š9∤•"Jºe—‹Œ @œvꩧ5*¤Zf‰—-o:¹H àê«°ºZ—™Ù©'BåÁ¼öêA‡è½¤£ŽT„-3£Nùۍ8B¶‡ÅAô¢ %ÛH|Ó@µÉžD¥c¨-ˆ]‹ÿý‚­›"pP@)jW=ÛPp†€µ‹TÄÑaî*;‚§:+Y®Â6x" •¶$ôëí²F:e¥¸ü±öÔÿqð9ð$Ó'Æ3v^Cë SÜ%‡€&`pNI'ŸC5ð'ÇÁfÌrø Òñ 3Ú3'þÞÇžƒÖÝL°Î˜...Xìm[]À ô'z¥ŽòF‹ päΨ8J‰ëÈ£À/U} ''Üa¸{€Óˆ"€€Q/@wÈæÝئÀ«Y5S%dËÕG#€Á "0@Ä"l"¾‹sÖ9RuÕDŒ0W-ÓS×'û...,...¸žê¶Å'nN':Ÿg„èe(à@Ã"ÀÂ>×ô¾ÔŽÜ¹ðÛLT%ê5;¢u*ðÙ34ð‚ÔWoý ¸@ÊöÛ @µ!p/þÿ‚0L>ý¤Ì¦Ì„¥Ïùmˆü¡•€Á€

            {òþÿ (šTÈÏÄOhX›éIaÀƁÜÀˆÀ NXKƒ"NðÌ 4˜ ,°T E¡„ `€½þiBz=˜ÀH:æÂPÂÇû@ t0øˆ<¬Vp‚ ø@Fa©P-z°UÁ7Ô€=¨A,@ °à‡€Á@ÂÆí :È%ªHÆŽE‡þKDr0ì`9A]@ Àà;à P ¨Á‰ƒìÀ9ðArIBÿ'...t(€Eú98X°¬`.œ@Z ƒà 4a & Fî 8häœÈð€˜¬$ }

            À§ì@/£jƒH `ô~0LT¬@0¢h°VÀ™3XA&u°‚T 6ÐZ9à9¨ R¹ÊÀ1.ØÁv0ƒ €ù"n 0³hЃL€ñœÁp ƒè@„rÔã3KNy-³ÖŒg&°¼€:€YIs š˜€$gÐk¶`$jZ IVà9à<ï9€î± @è l Â[²+ÿøåäÓ|rT£u¦TE¼ Ží„Af@ØàÖ 8úFD¾Q*f t0ƒQR@Ù„«zpÍ´µqÌÁ5ÅØE€±-́xéÇ ´à'0°$ ¨¸ƒÎNÀ"ð6‡YöÀ:ø(">ðƒÔ@-¨A pp¼àtA Îhà@.¸Á,ÐÔ'ðAl]ÛÜ6- ³Ý-n]`7{ÀÁnq`XàgôÀkpƒ¸º$! |P°` @¡m°[è bZ% xøõ‡P`1#þX0þ¯ æÙ™ðÁ„øÁ `C[øÂÞD ;

            See bug listed at: https://support.atlassian.com/browse/CSP-15544

            Pam Cole added a comment - We are still encountering this error several times an hour, even after upgrading to Resin 3.1.4: Receive error messages similar to the following after clicking Edit or Add New Page: Ézn°Ý¸) [Þ„|ÿ3ý8uÝA1©LÜØÚ''vC½.u‡-‡K|P8 ä ЯJºÒU_t- }ˆ ïÁ}FÙÐRùQ¦ž] §¡wÈ%ÀÑØXqú Sµ`€³Qˆ...çågP&ôR‹ô! & žˆ™(zMD7­=ø ÌG£&6:ÖšCV¡Ö"ÿï©PÛŒCFW^irUåÔ'X¶Ä$ +0bO`†)&˜ X' ÷Pp&! 0Â;ÔY'vÚé" ÙfXZùЛq.2'2ˆ&ÊPðÉ š9∤•"Jºe—‹Œ @œvꩧ5*¤Zf‰—-o:¹H àê«°ºZ—™Ù©'BåÁ¼öêA‡è½¤£ŽT„-3£Nùۍ8B¶‡ÅAô¢ %ÛH|Ó@µÉžD¥c¨-ˆ]‹ÿý‚­›"pP@)jW=ÛPp†€µ‹TÄÑaî*;‚§:+Y®Â6x" •¶$ôëí²F:e¥¸ü±öÔÿqð9ð$Ó'Æ3v^Cë SÜ%‡€&`pNI'ŸC5ð'ÇÁfÌrø Òñ 3Ú3'þÞÇžƒÖÝL°Î˜...Xìm[]À ô'z¥ŽòF‹ päΨ8J‰ëÈ£À/U} '' Üa¸{ €Óˆ"€€Q/@wÈæÝئÀ«Y5S%dËÕG#€Á "0@Ä"l"¾‹sÖ9RuÕDŒ0W-ÓS×'û...,...¸žê¶Å'nN':Ÿg„èe(à@Ã"ÀÂ>×ô¾ÔŽÜ¹ðÛLT%ê5;¢u*ðÙ34ð‚ÔWoý ¸@ÊöÛ @µ!p/þÿ‚0L>ý¤Ì¦Ì„¥Ïùmˆü¡•€Á€ {òþÿ (šTÈÏÄOhX›éIaÀƁÜÀˆÀ NXKƒ"NðÌ 4˜ ,°T E¡„ `€½þiBz=˜ÀH:æÂPÂÇû@ t0øˆ<¬Vp‚ ø@Fa©P-z°UÁ7Ô€=¨A,@ °à‡€Á@ÂÆí :È%ªHÆŽE‡þKDr0ì`9A]@ Àà;à P ¨Á‰ƒìÀ9ðArIBÿ'...t(€Eú98X°¬`.œ@Z ƒà 4a & Fî 8häœÈð€˜¬$ } À§ì@/£jƒH `ô~0LT¬@0¢h°VÀ™3XA&u°‚T 6ÐZ9à9¨ R¹ÊÀ1.ØÁv0ƒ €ù"n 0³hЃL€ñœÁp ƒè@„rÔã3KNy-³ÖŒg&°¼€:€YIs š˜€$gÐk¶`$jZ IVà9à<ï9€î± @è l  [²+ÿøåäÓ|rT£u¦TE¼ Ží„Af@ØàÖ 8úFD¾Q*f t0ƒQR@Ù„«zpÍ´µqÌÁ5ÅØE€±-́xéÇ ´à'0°$ ¨¸ƒÎNÀ"ð6‡YöÀ:ø(">ðƒÔ@-¨A pp¼àtA Îhà@.¸Á,ÐÔ'ðAl] ÛÜ6- ³Ý-n]`7{ÀÁnq`XàgôÀkpƒ¸ º$! |P°` @¡m°[è bZ% xøõ‡P`1#þX0þ¯ æÙ™ðÁ„øÁ `C[øÂÞD ; See bug listed at: https://support.atlassian.com/browse/CSP-15544

            Chris Chen added a comment -

            I believe this bug has been fixed and is closed on the resin side. Unfortunately, I am unable to test whether the fixed is done. It is fixed in 3.1.4

            Chris Chen added a comment - I believe this bug has been fixed and is closed on the resin side. Unfortunately, I am unable to test whether the fixed is done. It is fixed in 3.1.4

            Matt Ryall added a comment -

            Any more details on how this happens, Scott, or what the fix was? Your bug doesn't provide any further details:

            http://bugs.caucho.com/bug_view_page.php?bug_id=2131

            If possible, we'd like to add a test to our suite of Resin functional tests to ensure we can reproduce this problem in future.

            Matt Ryall added a comment - Any more details on how this happens, Scott, or what the fix was? Your bug doesn't provide any further details: http://bugs.caucho.com/bug_view_page.php?bug_id=2131 If possible, we'd like to add a test to our suite of Resin functional tests to ensure we can reproduce this problem in future.

            Thanks for that Dan. Good luck with the move to 3.1 if you get there.

            Paul Curren added a comment - Thanks for that Dan. Good luck with the move to 3.1 if you get there.

            As you suggest, replicating the problem is quite a challenge. We are currently on Resin v3.0 and v3.1 isn't a trivial upgrade – we are deploying to a new structure which might offer an opportunity, however the temptation of simply moving onto Tomcat v5.5 might be too high to resist.

            I'll let you know if we deploy and replicate the fault.

            Dan Hardiker added a comment - As you suggest, replicating the problem is quite a challenge. We are currently on Resin v3.0 and v3.1 isn't a trivial upgrade – we are deploying to a new structure which might offer an opportunity, however the temptation of simply moving onto Tomcat v5.5 might be too high to resist. I'll let you know if we deploy and replicate the fault.

            Thanks for that update Scott.

            Hopefully some of the watchers of this defect who can reproduce the problem will be able to confirm a fix in 3.1.4 once it's out there.

            Paul Curren added a comment - Thanks for that update Scott. Hopefully some of the watchers of this defect who can reproduce the problem will be able to confirm a fix in 3.1.4 once it's out there.

            This looks like a Resin issue. We're releasing a fix in 3.1.4 this week.

            The issue is very specific to the exact length of the result payload (exactly 16k bytes), so I'd expect it would be very difficult for you to reproduce.

            Scott Ferguson added a comment - This looks like a Resin issue. We're releasing a fix in 3.1.4 this week. The issue is very specific to the exact length of the result payload (exactly 16k bytes), so I'd expect it would be very difficult for you to reproduce.

              Unassigned Unassigned
              ivan@atlassian.com Ivan Benko [Atlassian]
              Affected customers:
              2 This affects my team
              Watchers:
              4 Start watching this issue

                Created:
                Updated:
                Resolved: