چند روز پيش متن زير را در سايت گذاشته بودم. به آن متن حرف ديگري نمي توان اضافه کرد، بجز اينکه اينطوري و با اين امکانات نمي شود سايت داشت. نوشته شده بود:
سلام. از ديشب تا لحظاتي پيش سايت ديدگاه بدون اطلاع قبلي فعال نبود. در اين مدت شما دوستان عزيز با ارسال ايميل و تماس تلفني هم مرا از قطع بودن سايت ديدگاه مطلع کرديد و هم با ابراز لطف رهنمودهاي دلگرم کننده داديد.
و اما چرا، ديدگاه هر چند وقت يکبار به اين درد دچار مي شود را قبلا توضيح داده بودم – ديدگاه به همان اندازه که پول بدهد آش مي خورد. اشکال مهمتر بر سر اين است که ما مي خواهيم کم پول بدهيم و آش زياد بخوريم، و نتيجتا شرکتهاي سرويس دهنده را با معضل روبرو مي کنيم. آنها انتظار چنين حجم از کارکشي از سرورهايشان را ندارند. سرورها هم در مواردي به صدا در مي آيند. براي اينکه گزارش فني بهتري داده باشم، گزارش شرکت سرويس دهنده را در زير مي آورم.
طبق معمول قول داده اند که ديگر چنين مساله اي پيش نيايد، امکانات را بهتر کرده اند، محدوديت ها را حذف کرده اند، و... ما هم طبق معمول قول آنها را مي پذيريم – چارهء ديگري هم نداريم.
با سپاس از شما
علي ناظر – 7 اسفند 1385
گزارش شرکت سرويس دهند:
Thank you for bringing this matter to our attention. We have been hard at work looking into the Apache failure on this server. In the late evening of yesterday, Sunday the 25th, the Apache web service had failed on this server. While the server was still online and operational, the Apache service (which handles all website traffic) was unavailable. The specific error pertaining to this problem involved the Apache service hitting the maximum amount of file handles permitted to be opened by a given system process. Despite various attempts to increase this limit (both the server's soft and hard cap), the problem remained. The issue at hand was simply that, while our server is *more* than capable to handle such a file handle load, the server's kernel was limiting the file handles to a default value.
After much investigation and Apache recompiling we were able to determine the original cause for the sheer amount of file handles that Apache was opening. As mentioned, our server is more than powerful enough to handle the level of file handles that Apache was attempting. The server was operating within optimal server performance ranges. The limitation of file handles per-process placed by the system kernel is something that would take significantly more downtime to modify. We are very pleased to have located the original cause as to eliminate the need for such further downtime.
At this time the server is running without any problems and within optimal levels. You should not expect any further problems with this server. Having located the origin of the issue, we can assure you such a problem will not reoccur. If you experience any anomalies with your site, please reply to this ticket so that we my immediately investigate and resolve the matter.
As mentioned, the only service affected was the Apache web server daemon. This means that the mail server, ftp server, and all other services on this server were in full operation throughout this issue. Any mail sent during that time will have delivered properly and was unaffected by the problem experienced with Apache.
I thank you for your patience during this outage and please accept our apologies for the inconvenience. An outage as lengthy as was experienced is far from acceptable. We are confident that as a result of our investigations we can assure this issue will not reoccur in the future.
If there is anything that I can do to ensure that this matter is fully resolved to your utmost satisfaction, please do not hesitate to reply to this ticket. I will be happy to help.
- Brian O.
Assistant Technical Manager